THE UNIVERSITY 0# ILLINOIS LIBRARY 630.6 VA2- cop.3 This book has been DIGITIZED and is available ONLINE. ^ JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY VOLUME 12 1920 PUBLISHED BY THE SOCIETY PRESS OF THE NEW ERA PRINTING COMPANY LANCASTER, PA. DATES OF ISSUE. Pages 1-44, January 20, 1920. Pages 45-/6, February 16, 1920. Pages 77-116, March 21, 1920. Pages 1 17-148, April 22, 1920. Pages 149-184, May 20, 1920. Pages 217-240, December 1, 1920. ERRATA. In Vol. ii, p. 335, 1. 18, read " 1,750,000 tons " instead of " 1,750,- 000,000 tons." In Vol. 12, No. 1, cover, title of paper by John H. Parker should read " A Preliminary Study of the Inheritance of Rust Resistance in Oats." ' In Vol. 12, p. 195, 1. 13, read "in Lychnis (14)'' instead of " in Lychnis (15) " &>3£, (o * v. i 2j >op. b / // 1 . CONTENTS, l] No. i. JANUARY. Page. Mooers. C. A. — Planting Rates and Spacing for Corn under Southern Conditions (Figs. 1-3) 1 Parker, John H. — A Preliminary Study of the Inheritance of Rust Re- sistance in Oats (PI. 1, 2) 23 ^ Richey, Frederick D. — Formaldehyde Treatment of Seed Corn 39 Agronomic Affairs. Membership Changes 44 Notes and News 44 i No. 2. FEBRUARY. Warburton, C. W. — Federal Seed Grain Loans 45 Buckman, H. O. — The Teaching of Elementary Soils 55 Smith, R. S. — Introductory Courses in Soils 58 Conner, S. D. — The Effect of Zinc in Soil Tests with Zinc and Gal- vanized Iron Pots 61 Fippin, Elmer O. — The Trufast Test for Sour Soil 65 Westover, H. L., and Garver, Samuel. — A Cheap and Convenient Ex- perimental Silo 69 Agronomic Affairs. Membership Changes 72 Notes and News 73 No. 3. MARCH. Jones, D. F. — Selection in Self-Fertilized Lines as a Basis for Corn Improvement (PI. 3-7 and Figs. 4-6) 77 Day, James W. — The Relation of Size, Shape and Number of Replications of Plats to Probable Error in Field Experimentation 100 Myers, C. H. — The Use of a Selection Coefficient (Figs. 7-9) 106 Scott, John M. — Bahia Grass 112 Agronomic Affairs. Membership Changes 114 Notes and News 115 No. 4. APRIL. Fippin, Elmer O. — The Status of Lime in Soil Improvement 117 Gaines, E. F. — The Inheritance of Resistance to Bunt or Stinking Smut of Wheat 124 Waldron, L. R., — First Generation Crosses Between Two Alfalfa Species (Figs. 10-13) 133 537040 vi CONTENTS. Page. Sieglinger, John B. — Temporary Roots of the Sorghums 143 Agronomic Affairs. Membership Changes 145 A New Committee on Varietal Standardization 146 Notes and News 147 No. 5. MAY. Briggs, Glen. — Guam Corn 149 Harris, F. S., and Butt, N. I. — The Unreliability of Short-Term Experi- ments 158 Spragg, Frank A. — The Coefficient of Yield 168 Carrier, Lyman. — The History of the Silo 175 Agronomic Affairs. Membership Changes 183 Notes and News 183 No. 6-7. SEPTEMBER-OCTOBER. Richey, Frederick D. — The Inequality of Reciprocal Corn Crosses (Fig. 14) 185 Brown, Ernest B. — Relative Yields of Broken and Entire Kernels of Corn 196 Wentz, J. B. — An Outline of an Undergraduate Course in Grain Grading (PI. 8) 198 Harlan, Harry V. — Smooth-Awned Barleys (Fig. 15) 205 Agronomic Affairs. Annual Meeting of the Society 209 Notes and News • 209 Conference on Elementary Soil Teaching 211 Conference on Western Agronomists 214 Membership Changes 216 No. 8-9. NOVEMBER-DECEMBER. Harris, Frank S. — The Agronomist's Part in the World's Food Supply (Presidential Address) • 217 Agronomic Affairs. Membership Changes 225 Notes and News 226 Report of the Secretary-Treasurer 227 Minutes of the Thirteenth Annual Meeting 229 Report of Committee on Standardization of Field Experiments 233 Report of Committee on Terminology 233 Report of Committee on Varietal Standardization 234 Report of Advisory Board to National Research Council 235 Report of Editor 237 Index 238 JOURNAL OF THE American Society of Agronomy Vol. 12. January, 1920. No. 1 PLANTING RATES AND SPACING FOR CORN UNDER SOUTHERN CONDITIONS. 1 C. A. Mooers. Statement of the Problem. In Tennessee, as in other Southern States, a wide range of varie- ties of field corn is grown. Some are recognized as standard in the corn belt ; for example, Reid Yellow Dent and Boone County White. Others are standard in the South, such as Hickory King and the prolines — Albemarle, Cocke, Biggs, Mosby, etc. Varieties differ ma-- terially with regard not only to length of growing season but also to height of stalk and to foliage production. These differences are espe- cially marked when the varieties are grown on rich land. On alluvial land at the University farm, Knoxville, Tenn., when Huffman grows 12 feet tall Learning will reach a height of about 8 feet. It is not surprising, therefore, to find that on the same kind of land different varieties require different rates of planting in order to produce both the largest yield and the best quality of grain. Also, as has been pointed out previously, 2 varieties of similar length of season and habit of growth may differ appreciably in the rate of planting which gives best results. A wide variation in productivity is found among Tennessee soils. On the same farm it is not uncommon to find uplands which produce 1 Contribution from the Tennessee Agricultural Experiment Station, Knox- ville, Tenn. Revision of a paper read at the eleventh annual meeting of the American Society of Agronomy, Baltimore, Md., January 7, 1919. Received for publication July 27, 1919. 2 Mooers, C. A., Stand and soil fertility as factors in the testing of varieties of corn, Tenn. Agr. Expt. Sta. Bui. 89, p. 49. 1910. 2 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. only 20 bushels of corn per acre' and lowlands which produce 50 bushels or more. Common experience has shown that these differ- ent soils should be planted at different rates and, as a matter of fact, the corn on rich land is planted relatively thick and on poor land relatively thin, the rate being determined in every case by the judg- ment of the grower. Tho experienced corn growers are probably not far wrong when planting a well-known variety on a well-known soil, a material difference of opinion frequently exists, as might be ex- pected, regarding the proper stand under any given condition. At best it is a guess. Unfortunately, many men have poor judgment in the matter, as can be seen at the close of a favorable season, a too thick stand being often in evidence. A reliable guide is therefore needed in order to determine, with some degree of accuracy better than a mere guess, the proper stand of corn with regard to both the productivity of the soil and the variety to be grown. In this connection the best rate of planting for silage is of interest, as is also the effect of variation in the grouping and spacing of a given number of plants. The Rate-of-Planting Expeeiments and the Determination of the Varietal Factor. the experimental rates of planting. In the varietal trials conducted in various parts of Tennessee, the varieties of corn have for several years been planted in three blocks, each at a different rate. On the rich bottom land at the Knoxville station the rates have been 6,000, 8,000, and 10,000 plants per acre. On upland of good quality the rates have been 4,200, 5,400, and 6,600 plants. At the Jackson station in West Tennessee and in the Middle Tennessee experiments the rates have been 3,000, 4,200, and 5,400 plants. Soils differing considerably in productivity have been used, but in general the yields have been good, ranging from 30 to 80 bushels per acre. However, some experiments have been made on poor land, which under average conditions produced less than 30 bushels per acre. Data have been obtained which bear on the problem of the proper stand for each of a number of varieties of corn under various degrees of productiveness. The question then arose as to the possibility of assembling these data in such a way as to make the re- sults of general application under similar climatic conditions. Hickory King, as a rule, yielded most in the rich-land experiments when planted at the 10,000 rate and in the Jackson experiments at either the 4,200 or 5,400 rate. But how were the results to be placed on a MOOERS : CORN PLANTING RATES. 3 simple basis which would allow the two series to be compared and perhaps lead to their application elsewhere? RESULTS OBTAINED WITH HICKORY KING CORN. The results obtained with Hickory King corn will be taken as an example of the methods followed with other varieties. Table I Table i. — Yields in bushels per acre of Hickory King corn at various rates of planting, with average weights of grain per plant. Rate of planting. Selected rate. Location. Year. 6,000 plants. 8,000 plants. 10,000 plants. Yield per acre. Yield per plant. Yield per acre. Yield per plant. Yield per acre. Yield per plant. No. of plants. Yield per plant. Ivnoxvillc, bottom land 1906 a 1906 6 I907 a 1907 b 1 908 1909 I9IO I9II 1912 1913 1914 IOI ^ Bu. 64.I 37-4 49.2 68.5 42.8 64-3 50.4 66.4 68.9 42.9 55-2 72.9 Lb. 0.60 0-35 0.46 0.64 0.40 0.60 0.47 0.62 0.64 0.40 0.52 0.68 Bu. 8l. I 44-7 53-5 76.0 59-5 65.5 62.1 76.0 72.0 47.1 55-2 67.1 Lb. 0-57 0.31 0-37 o.53 0.42 0.46 0-43 0-53 0.50 0-33 0-39 0.47 Bu. 89.9 64.4 60.0 72.8 58-4 70.7 76.I 76.O 75-o 60.0 49.9 81.4 Lb. O.50 O.36 0-34 O.41 0-33 O.40 0-43 0-43 O.42 0.34 0.28 O.46 10,000 10,000 10,000 9,000 9,000 9,000 10,000 9,000 9,000 10,000 7,000 10,000 Lb. 0.50 0.36 0.34 0.46 0.37 0.42 0.43 0.47 0.46 0.34 0.44 0.46 56.9 0-53 63-3 0.44 69.6 0-39 9-333 0.42 4,200 plants. 5,400 plants. 6,600 plants. Knoxville, upland .... I9U 1915 TOTfS 39-9 35-3 34-0 0-53 0.47 0-45 49.O 37-2 46.0 0.51 o.39 0.48 52-7 37-2 42.0 0-45 O.32 O.36 6,000 4,800 6,000 0.48 0.43 0.41 36.4 O.48 44-1 0.46 44.0 O.38 5.6oo 0.44 3,000 plants. 4,200 plants. 5,400 plants. I9IO I9II 1912 1913 1915 I9l6 41-5 21.9 36-0 34-5 36.8 34-o O.74 O.41 O.67 O.64 O.69 O.63 43-3 30.2 41. 1 41.6 49.I 48.8 0.58 0.40 0-55 0-55 0.65 0.65 43-3 23.I 45-6 39-6 50.7 52.7 0-45 O.24 O.47 O.41 0-53 0-55 4,800 4,200 4,800 4,800 4,800 4,800 0.51 0.40 0.51 0.47 0.58 0.59 34-i O.63 42.4 0.56 42.5 O.44 4.700 0.51 I9l6 1917 1917 I9I8 I9I8 26.4 37-9 39-7 44.6 25-7 O.49 O.71 O.74 O.83 O.48 35-o 37-6 36.8 43-7 25-7 0.47 0.50 0.49 0.58 0-34 35-7 43-5 39-6 42.8 25-7 0-37 0-45 O.41 O.44 O.27 4,800 5.400 4,800 4,800 3.600 0.41 0.45 0.46 0.51 0.40 McMinnville Algood Crossville Average 34-9 O.65 35-8 0.48 37-5 0-39 4,680 0.45 a Early planting. 6 Late planting. 4 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. shows the yields of grain in 12 trials on bottom land and in 3 trials on upland at the Knoxville station, in 6 trials at the Jackson station, and in 5 trials in Middle Tennessee. 3 Along with every yield is given the average weight of grain per plant. In all trials only one plat, which consisted of two 3-foot rows 60.5 feet long, or one one-hun- dred-and-twentieth of an acre, was used for each rate of planting, but the location of the plats was changed from year to year. Relation Betzveen Yield per Acre and Weight of Grain per Plant. In the bottom-land trials at Knoxville the yields are decidedly in favor of the 10,000 plants per acre, the average grain production from that rate being 69.6 bushels. The next highest average is 63.3 bushels, from the 8,000 rate, and the lowest is 56.9 bushels, from the 6,000 rate. The weight of grain per plant is in the reverse order, the average being 0.39 pound from the 10,000 rate, 0.44 pound from the 8,000 rate, and 0.53 pound from the 6,000 rate. In the upland series, the average yield at both the 5400 and 6,600 rates was practically the same. The average weights of grain per plant were 0.48 pound at the 4,200 rate, 0.46 pound at the 5,400 rate, and 0.38 pound at the 6,600 rate. In the seven years' trial at the Jackson station the 4,200 and 54°° rates were tied for first place, with an average of 42.5 bushels of grain per acre, while the yield at the 3,000 rate was only 34.1 bushels. The average weights of grain per plant were 0.63 pound from the 3,000 rate, 0.56 pound from the 4,200 rate, and 0.44 pound from the 5,400 rate. In the Middle Tenessee trials the highest average yield was from the 5,400 rate, with 37.5 bushels per acre and 0.39 pound of grain per plant. The next highest yield was from the 4,200 rate, with an average of 35.8 bushels per acre and 0.48 pound of grain per plant. The lowest yield was from the 3,000 rate, which produced 34.9 bushels per acre, with an average of 0.65 pound of grain per plant. Taking the four series as a whole, it is evident that where the stand was thin so that the average weight of grain per plant was from 0.46 to 0.65 pound, the yield per acre was appreciably less than where the stand was so thick that only 0.38 to 0.46 pound of grain was pro- duced per plant. In other words, the data indicate the conclusion that where Hickory King is planted at such a rate that the average 3 Similar data relative to experiments with other varieties were submitted to the editors, but are not presented here because of the expense of printing. The data are summarized in Table 2. MOOERS I CORN PLANTING RATES. 5 weight of grain per plant is 0.38 to 0.46 pound the stand may be con- sidered as approximately right for the soil and season. Inspection of the results shows that the best number of plants per acre varies appreciably with the season. If the best rate, from a conservative point of view, is selected for each season and the weight of grain per plant is obtained, will not the average arrived at in this way be more closely correlated with the largest yield than the aver- ages previously discussed? Believing this to be true, the writer has entered these data in the last two columns of Table 1 under the head- ing of " Selected rate." In the selection of this rate and in the calculation of the weight of grain per plant to go with it in Table 1 and in similar data for other varieties, use was made of the following rules : 1. The highest yield determined the rate and was used as the basis of calculation unless the next highest was near enough to be con- sidered as a duplicate. 2. When two rates give rise to duplicate yields, the average of the two rates was taken along with the average of the two yields for the calculation of the factor. Duplicate yields are delimited as follows : a. Up to 40 bushels per acre, two yields were considered as dupli- cates if the difference was not more than 4 bushels. b. For yields from 40 to 60 bushels, two yields were considered as duplicates if the difference between them was not greater than 10 percent of the higher yield. c. With one or both yields greater than 60 bushels, two yields were considered duplicates if the difference between them was not greater than 6 bushels. 3. When the yields either of all three or of the lowest and highest rates were within the limits as given under paragraph 2, the two highest rates of planting were averaged if the yields were above the point of gravity, as determined by the medium rate of planting and the application of the finally accepted factor, 4 or standard weight of grain per stalk for the variety in question. With yields below the point of gravity, the two lowest rates are to be used. In either case the two highest yields were averaged as a basis for the calculation of the weight of grain per plant. This involves a " cut and try " method of procedure, but only for the small number of doubtful cases. Continuing with Hickory King as an example, the highest yield 4 The word " factor " is here used in its mathematical sense of a number entering into an equation to form a product. It is used as a substitute for the expression, " standard weight of grain per stalk." 6 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. for the early-planted crop of 1906 in the bottom-land series was 89.9 bushels at the 10,000 rate and, as the next highest yield could not be considered as a duplicate, the selected or probably best rate was 10,000 and the average weight of grain per stalk, 0.50 pound. In the case of the late-planted crop of 1907, the 8,000 and 10,000 rates gave rise to duplicates, hence the average of the two rates was taken and the average of the two yields was used in the calculation of the grain per plant at the selected rate. In the upland series for 191 5 the yields were, within the limit of error, the same for all three rates. The average of the two highest, 37.2 bushels, was taken as the yield, but the rate of 4,800, or the average of the two lowest, was taken as the selected rate, because it gave a weight of grain per stalk more in harmony with the varietal standard of 0.45 than either the 5,400 rate or an average of the 5,400 and 6,600 rates. According to the averages of the selected rate, the best number of plants in the bottom-land experiments was 9,333 and the average weight of grain per plant was 0.42 pound. In the upland experi- ments the averages were 5,600 plants and 0.44 pound of grain. In the Jackson experiments the averages were 4,700 plants and 0.51 pound grain. In the Middle Tennessee experiments the averages were 4,680 plants and 0.45 pound of grain. Using the selected-rate calculations, the average weight of grain per plant for all trials is 0.45 pound, and this number is taken as the standard for Hickory King. THE VARIETAL FACTOR. The weight of grain per plant most in harmony with the best yield, as calculated by the method outlined, will be referred to as the factor for the variety in question. Table 2 gives a list of varieties with standard or finally accepted factors as worked out either in this way or with a slight modification, as explained later. Average Expectancy Calculations. The total number of trials made with Hickory King is 26, which may be considered a fair number for a safe average. However, with the exception of Albermarle Prolific the number of trials for any of the varieties reported in Table 2 is considerably below 26, ranging from 5 to 19, so that the actual average may not be the best guide obtainable from the data. For example, the factor for the 1908 crop of the Little Willis variety is 0.39, which is nearly 25 percent less than the lowest factor in the 8 other trials. The factor 0.39 is in- cluded in the average for the bottom-land series, but is omitted in MOOERS : CORN PLANTING RATES. 7 Table 2. — Standard varietal factors, or average yields per plant producing the highest acre yields of corn. Average expectancy basis. Total Average v • anety. Number Factor trials. obtained. of trials selected as averaged. standard. Albemarle Prolific 28 0-573 28 0-573 Batt Prolific 6 .567 6 •567 16 .480 16 .480 16 • 576 16 .576 7 .630 6 •585 Hickory King 26 .448 26 .448 Hildreth 7 •546 5 .588 19 .617 14 .644 10 .387 10 .387 12 •538 12 .538 Kansas Sunflower 6 •458 5 .476 Learning 7 •373 5 .408 6 .462 6 .462 Lewis Prolific 17 .606 13 •54 2 Little Willis 9 •557 8 •578 Looney 11 .532 11 •532 McAuley 5 .428 4 .460 7 .496 4 .548 6 .523 6 .523 18 .588 15 .548 17 •458 17 •458 12 •493 12 •493 6 •523 6 .523 No. 182 13 .518 11 .496 the average expectancy calculation shown in Table 2. In a similar manner, factors which vary widely from the others in the same group are included in the series averages, but are omitted in the calculation of the average expectancy averages of Table 2. In all such cases the average expectancy is assumed to give the better figure. It may be noted that for 11 out of the 24 varieties for which standards are given in Table 2, the average expectancy factors are different from the actual average of all results. For the other 13 varieties no results were excluded, so that the actual average and the average expectancy average are the same. Formula for the Calculation of the Proper Stand of Corn. The yield of grain per acre may be calculated by multiplying the number of plants by the average production per plant. The number of plants is therefore equal to the yield in pounds divided by the average weight of grain per plant. If the yield is given in bushels, the equation is as follows : 8 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. where N stands for the number of plants per acre, Y the yield in bushels, and F the average weight in pounds of grain per plant, 56, of course, being the standard weight of a bushel of corn. If F, as the varietal factor, is worked out experimentally and the average yield or expectancy of a given field is known, the proper number of plants to be allowed is easily calculated. For example, if Hickory King, with an assumed factor of 0.448, is to be grown on land which in a fair season produces about 40 bushels per acre, how many plants should be allowed per acre? N = 40 X 56 0.448, 40 X 56 = 2,240, 2,240 -r- 0.448 = 5,000 plants. As the majority of seasons are favorable to a fair crop the assump- tion is made that no one should plant for the exceptional crop, either good or bad. It is probable that no good farmer does so in practice. Land that in a fair or average season produces 40 bushels of corn per acre should be planted therefore with that yield in view and not for either the occasional partial or the occasional extra large crop. This 40-bushel rate of planting does not mean, however, that little or no more can be produced in the extra favorable season. Examina- tion of Table 1 shows that average weights of 0.60 to 0.82 pound of grain per stalk are not uncommon. Therefore, a field planted for a 40-bushel expectancy may make, so far as the number of stalks is concerned, 50 to 70 or more bushels per acre. On the other hand, the man who plants 40-bushel land for a 60-bushel crop need not be sur- prised at a yield reduced appreciably during a normal season in both quantity and quality as compared with that obtainable with the proper, or 40-bushel, stand. Fortunately the expectancy is generally well known ; that is, the farmer knows whether a given field will, in a fair season, produce in the neighborhood of 20, 30, 40 or more bushels to the acre. As a matter of fact, in practice he keeps this expectancy in mind, planting thick or thin, depending on his judgment regarding the capacity of the soil. The uncertainty in the equation is the value to be given F. Is F a constant or, as seems highly probable from evidence given later, is it a variable, which may within certain limits approach a constant ? Effect of Variation in Season and in Productivity on the Varietal Factor. If factors obtained under conditions of high yield show a constant difference from those obtained under conditions of low yield, the MOOERS : CORN PLANTING RATES. 9 formula as given is evidently incomplete. The subject could logically be considered under two heads : ( i ) Variation due to differences in season and (2) variation due to differences in soil productiveness. 5 Under the former head a soil of constant high productiveness is required, a condition which is well met by the bottom-land series. To meet the requirements of the second head, constant seasons and variable soils would be required. As this condition was not obtain- able, the data from the different localities are considered as a whole and with the understanding that several variables influence the results. INFLUENCE OF VARIATION OF SEASON. The bottom-land series, which was conducted on a strong, rich soil and which was continued for the longest time and with the largest number of varieties, gives some valuable data as to the effect of season on the varietal factor. Table 3 is arranged to show the factors obtained from different yields for 19 varieties. Owing to the limited data for any one variety, stress should be laid on the averages rather than on the results for a single variety. Two averages from each yield are given. One is the actual average and the other, or corrected average, was obtained from the actual by making allowance for dif- .60 />CToH .55 .SO .14-0 .as 36" ±5 SS CS 7S- 8& fS wB'vt Fig. 1. Curve showing effect of season on the varietal factor. 5 As all the data were obtained under similar climatic conditions, no attempt is made to discuss the effect of climate. IO JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. ferences chargeable to the blank spaces which give rise to averages from different sets. This correction, which was in general of minor importance, was accurately made by making use of the average factors as obtained in the bottom-land series. Figure I shows the curve obtained by plotting the corrected averages as ordinates and the yields as abscissas. Table 3. — Influence of season on the varietal factor as shown on results on bottom land at Knoxville. a Yield in bushels per acre. Variety. 40-49 (45). 50-59 (55). 60-69 (65). 70-79 (75). 80-89 (85). 90-99 195). Albemarle Prolific 0-37 •43 •34 O.46 2 -50 •43 3 .50 4 -37 2 o-5i 4 0-53 •56 2 -58 2.61 6 -45 •56 4 .62 2 -43 2.44 •49 •49 .61 2 -56 2.61 2 o.68 •76 3 .49 .66 2.48 2 - 50 3 - 53 Batt Prolific 4 -43 2 -52 •36 2 -56 .48 2.42 •52 •37 •37 2 -58 2 -5i 2.49 3 o.69 Hildreth 2.48 3 .32 •37 3 -34 3 -30 2 .62 •32 .48 .46 .46 •71 .28 .28 .46 Little Willis •52 •56 •50 •53 •53 •47 •51 •30 2.41 2.41 •47 .46 •45 2 -55 2 .63 4 -45 4 -50 •56 •75 Reid Yellow Dent •33 2.47 2 -4i •53 (2) corrected n -352 •371 29 -425 •434 27.467 .472 42 -540 •535 24.548 •543 8.630 .569 a The number of trials averaged, where more than one, is indicated 1 by the superior figures. The lowest average factor, as corrected, is 0.371, obtained at the average yield of 45 bushels per acre ; the factor at the 55-bushel aver- age is 0.434 ; at the 65-bushel average, 0.472 ; at the 75-bushel average, 0.535 5 at tne 85-bushel average, 0.543, and at the 95-bushel average, 0.569, or the highest of all. It is evident, therefore, that the amount of grain per plant increases with the favorableness of the season even when planted at the best rate for acre production. Attention is called to the fact that the average of the standard factors for these 19 varieties is 0.514, which is within 5 percent of the average of all the figures in Table 3. MOOEKS : CORN PLANTING RATES. INFLUENCE OF VARIATION IN SOIL PRODUCTIVENESS. Factors Obtained in Different Localities. Table 4 shows the varietal factors as obtained in different localities and Table 5 is a summary of the results of Table 4. From this sum- mary it is evident that the Jackson series gave the highest factors, which average II. 17 percent more than those obtained in the bottom- land series. The Middle Tennessee series gave factors which on the average were only 0.6 percent less than the bottom-land results. The upland series gave factors which averaged 6.43 percent less than the bottom-land series. Table 4. — Varietal factors or yield of com per plant as obtained in different localities. Variety. Kno: Bottom land. cville. Upland. Jackson. Middle Tennessee. Yield per acre. Factor. Yield peracre. Factor. Yield peracre. Factor. Yield peracre. Factor. Bu. 70.6 63 -9 67.4 69.6 71.4 68.6 56.8 67-5 62.4 67-3 78.9 79-3 66.6 70.6 68.7 75-8 O.58 .46 •58 .42 •54 •57 •37 •59 •53 •55 •44 .62 •44 .48 •55 .46 Bu. Bu. 42.3 43-2 36.5 42.5 50.2 45-8 40.9 45-5 43-7 40.1 45-8 56.1 40.1 37-0 O.65 •54 •56 •51 •57 •73 .46 .70 •59 •50 .61 •59 •47 •53 Bu. 41.0 O.46 Boone Co. White Cocke Prolific Hickory King Hildreth 44.1 O.44 36.7 •45 28.7 53-6 •34 .48 Lewis Prolific Little Willis 46.0 •50 5i-3 45-i 45-i 40.5 •55 •53 •58 •52 Mosby Prolific Neal Paymaster Reid Yellow Dent Webb Improved Watson . . Wild Goose 54-2 35-2 •55 .42 45-o 47-4 •50 .48 No. 182 46.4 •58 43-5 •54 Table 5. — Summary of data shown in Table 4. Locality. Number of varieties averaged. Yield per acre. Factor. Increase or decrease ( — ), as compared with bottom land factor. Bushels. Percent. 6 45-3 O.480 - 6.43 6 7i-3 •513 15 43-7 •573 11. 17 15 69.1 .509 9 43-9 .494 — 0.60 Knoxville bottom land 9 70.3 •497 12 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Locality Data Arranged on a Comparable Basis. In order to get an accurate comparison between factors obtained under different soil conditions, not only must the same varieties be grown under the same climatic conditions, but also the range in the rates of planting must be of the same order. In the four series under consideration there were variations both in seasonal conditions and in the range of the rates of planting. The latter is difficult to gauge perfectly, at least where only three rates of planting are used. In the series under consideration the lowest factors were obtained in the upland series where the rates of planting were the highest for the yield. They were probably too high for the conditions. The op- posite appears to have been the case in both the Jackson and the Middle Tennessee experiments, that is, the rates were low for the yields obtained and this may account, at least in part, for the rela- tively high factors in the Jackson series. The rates used in the bot- tom-land series seem to have been good for most varieties. In none of the series, however, do the rates appear to be seriously wrong, so that their average is considered to be at least reasonably satisfactory. Based on the assumption that the differences between the series when regarded as a whole were due to inequalities in the ranges of the rates of planting, to the brevity of the experiments, etc., the re- sults of both the Middle Tennessee and the Jackson series were raised and those of the upland series reduced to the basis of the bottom- land series. Table 6 gives the data from 15 varieties as thus prepared at 6-bushel intervals to show the effect of yield on the varietal factor. Figure 2 shows the curve obtained by plotting the corrected averages, which were obtained in the same way as for Table 3. ■40 1 | I I J I I I | J 27 33 31 4-5 SI 57 63 6f J 5 SI 87 2«< Fig. 2. Curve showing influence of soil productivity on the varietal factor. The broken line connects the points as found and the straight, solid line is the theoretical deduction. MOOERS : CORN PLANTING RATES. 13 Table 6. — Influence of soil fertility on the varietal factor, a composite from all series when placed on a comparable basis.' 1 Yield in bushels per Acre. Variety. 24-29. 3°-35- 36-41. 42-47. 48-53- 54-59. 60-65. 66-71. 72-77. 78-83. 84-89. Albemarle Prolific , 4 0.49 4 o.48 5 o-56 0.58 2 o-55 3 o-55 4 0-53 2 o.68 Boone Co. White .40 2 -43 2 -39 3.50 2 -38 0.38 2 -58 20.49 •50 Cocke Prolific 2 .6o 3 -58 •49 2 -50 •39 .64 • 72 2 -55 Hickory King o.37 2 -36 4 -4i 4 -43 3.47 4 -37 •36 .42 4 .46 .46 •50 Huffman . 3 -54 2 -54 .48 2 -50 .66 4 -63 3 -55 Iowa Silver- mine •37 2 -32 •43 2.32 2.41 2 -43 Jarvis Gold- en Prolific 2.46 *. 4 8 2 -59 .66 3.6o Lewis Pro- lific 2.49 •50 •75 5 -55 2.64 3 -59 .70 .61 • 71 Little Willis .58 2 -49 • j •39 .62 2.56 Looney 2 -56 3.52 .61 2.49 •47 2 -6 5 Mosby Pro- lific •39 2.44 •49 .42 .46 Neal Pay- master . . . •47 2.46 3 -54 •58 4.66 .64 2.72 2.70 •53 Reid Yellow Dent 2.40 3.44 4 -43 •45 •39 2.48 3 -45 •47 Webb Improved 2.46 •38 •52 •44 •39 4 -43 •38 .68 •49 •52 •52 No. 182 .41 6.47 •47 •47 Averages : Actual . . . 3.403 "•449 25 -47i 39 -505 27 -5i2 • 23.524 23.468 15 -543 29 -55i 13 -5i3 8 -566 Corrected . •455 •452 .485 .505 •504 •5ii .466 •533 •550 •515 • 577 a Superior figures show number of trials averaged, when more than one. As shown in Table 6, the greater part of the trials produced yields between 36 and 78 bushels per acre. Between these limits the cor- rected factors are perhaps as concordant as could be expected from the limited data, except for the very low factor of 0.466 in the 60-65- bushel column. The writer considers this column to be abnormal. However, taking the results as they stand, two interpretations seem possible ; one, that they indicate a constant factor between the pro- ductivity limits of, say, 36 and 66 bushels, and the other that, with an increase of productivity, there is an increase in the factor as indicated in figure 2. The latter view is strengthened by inspection of the data given in Table 7, which gives the results of rate-of-planting trials under conditions of low yield. Based on the course of the theoretical line shown in figure 2, the calculation was made that there was a variation of 0.0014 m the factor JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. for each bushel change in soil productivity. According to Table 2, the average of the standard factors for the 15 varieties is 0.521. This was found to coincide with a 64-bushel yield on the theoretical line. Taking the standard factors of Table 2 as standard for 64-bushel land, the change to be made for a soil of other productivity is easily calculated by means of the following formulas : 1. For land producing on the average more than 64 bushels per acre — N = 56 Y F + (F - 64).ooi4 2. For land producing on the average less than 64 bushels per acre — N 56F F - (64 - F).ooi4 The Varietal Factor Under Poor-Land Conditions. The data previously discussed include few obtained under poor- land conditions ; that is, land producing about 25 bushels or less per acre. Table 7 gives the best data in the possession of the writer with regard to the effect of poor land on the varietal factor. In the first section of Table 7 the average yield at each of the three rates is practically constant, and lower rates of planting were evidently needed in order to get the information desired. In the sec- ond section, the 3,200 rate gives the highest yield, but owing to the limited number of results little stress can be laid on the outcome. The comparatively slight effect of the variations in the rates of plant- ing on the yield thruout the series is of special interest, because it shows that there can be a considerable range in the rate of planting on poor land without serious consequences. According to Table 2, the average factor for the nine varieties rep- resented in the first section of Table 7 is 0.52 and for the four varie- ties of the second section is 0.51. Either average is higher than that calculated for any rate of planting in Table 7 except for the 2,000 rate, which gave an inferior yield. Altho the data indicate a reduction in the varietal factor for poor- land conditions, much more evidence is needed in order to find out the exact change that should be made. The writer considers it prob- able, however, that no land should be planted at a less rate than that calculated for a 25-bushel crop. MOOERS : CORN PLANTING RATES. I 5 Table 7. — Varietal factors under conditions of low yield. Yield in bushels per acre. Variety. Place. Year. 3,000 rate. 4,200 rate. 5,400 rate. Alheinarle Prolific Tullahoma 1916 16.8 20.3 22.5 Do 1VT a vl a n rl 1918 18.3 19.7 18.5 Hickorv Kin° r Jackson IOI A 23.2 20.8 13-2 Do Tullahoma I9l6 24.6 21.0 24.6 Do Baxter I9l8 16.3 17. 1 19.7 Do Crossville I9l8 25.7 25-7 25-7 Tarvi«! frolrlpn T^rnlifir* Tullahoma I9l6 22.4 18.9 21.0 Do Baxter I9l8 20.6 24.O 18.0 Do IVlayland I9l8 22.3 17. 1 23.2 Lewis Prolific Tullahoma I9l6 21.0 21.0 24-6 Looney Baxter I9I8 18.8 19.7 18.0 Do Crossville I9I8 2 5-7 30.0 29.1 TV K 1 1 ^. Marlboro Mayland I9l8 17-5 21.0 17. 1 Neal Paymaster Tullahoma I9l6 22.4 21.7 26.7 Baxter I9I8 22.3 25-7 22.0 Reid Yellow Dent Smithville 1912 19.1 23-4 24.8 No. 182 Tullahoma I9l6 19-6 10. 21.0 Do Baxter I9I8 21.4 26.6 2A.0 Mayland I9I8 18.8 Average vields 20.9 21.7 21.8 Factors from average yields . . •39 .29 .23 2,000 rate. 3,200 rate, 4,400 rate, Knoxville I9I8 23.2 27.6 27.6 McMinnville I9I8 15-4 25-4 23-4 Do. I9I8 25.O 25-7 23.6 No. 182 Do. I9I8 26.3 32.1 31.2 Average vields 22.5 ,7.7 26.5 Factors from average vields .63 .48 •34 A TOO HIGH VERSUS A TOO LOW RATE OF PLANTING. Two series covering much wider ranges of planting than those of previous years were carried out in the summer of 1918. One was conducted at the Jackson station, in West Tennessee, and the other at the Knoxville station, in East Tennessee. The data are given in Table 8 and the results are shown graphically in figure 3. The soil selected at the Jackson station was poor, with an expect- ancy of only 20 or 25 bushels in a fair season. The season of 1918 was, however, decidedly unfavorable on account of dry weather. The soil at the Knoxville station was of high productivity, capable of pro- ducing 60 or 70 bushels per acre in a fair season. The results from the two series agree in showing a rather rapid rise in yield with the increase in the number of plants, followed by a comparatively gradual decline after the maximum is reached. This i6 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. result is of interest as indicating that the margin of safety lies with a too high rather than a too low rate of planting. f/6 8 -fit /ooo jt.000 3oeo 4.000 Sooo boeo JOOo gooo fooo /ocoo 1/000 /2.000 T^LJj/vri Fig. 3. Curve showing results of rat'e-of-planting experiments with corn as obtained at Jackson (lower curve) and at Knoxville (upper curve) in 1918. Table 8. — Results from two series of rate-of -planting experiments during the season of 19 18. Yield per acre. Plants per acre. Jackson station." Knoxville station, 4 Grain, bushels. Stover, tons. grain in bushels. 1,000 9-8 O.32 15.2 2,000 14.6 •43 25-7 3,000 17.8 .62 29.6 4,000 13.2 •59 40.0 5»ooo 14.2 •71 46.5 6,000 14.O .80 50.5 7,000 9.9 .82 5i-7 8,000 50.8 9,000 46.7 10,000 41.2 11,000 43-5 12,000 37-8 a The soil on which the corn was grown at Jackson was poor and the season very unfavorable. The different rates were planted in fortieth-acre plats repli- cated four times, except the 2,000 and 7,000 rates, which were replicated only three times. 6 At Knoxville, the soil was good, but the season was below the average. All plats were duplicated. MOOERS : CORN PLANTING RATES. 1 7 RATE OF PLANTING, FOR SILAGE. Table 9 gives the total production of ear corn and stover in three series of rate-of-planting experiments. The yields were obtained by adding together the weights of stover when well cured in the shock and of ear corn when in a thoroly air-dry condition. In this table only those varieties were included which made their greatest average grain production at rates well below the maximum rate of planting. Table 9. — Yields per acre of combined ear corn and stover on a field-cured basis at different rates of planting. KNOXVILLE BOTTOM LAND. No. of Yields in pounds per acre. Best rate Variety. years. 6,000 plants. 8,000 plants. 10,000 plants. for grain yield. « Albemarle Prolific 13 9.769 10,067 IO,6l7 6,929 Cocke Prolific II 9.983 9,826 10,456 6,601 Hildreth 5 9,666 10,480 10,329 6,847 Huffman 11.433 11,284 12,048 5.965 Legal Tender 8,043 9,109 9.390 7.745 Lewis Prolific t 9.325 8,882 9,082 6,974 Little Willis 5 9,2ii 9,438 9.276 6,046 5 8,958 8,864 8,868 7,084 Average 7.8 9-549 9,744 10,008 6,774 KNOXVILLE UPLAND. 4,200 plants. 5,400 plants. 6,600 plants. Hickory King 3 5,068 6,080 5,960 5,673 Jarvis Golden Prolific 3 5-799 6,022 5,981 4,705 Neal Paymaster 3 6,152 6,954 6,86l 5,54o Wild Goose 3 6,086 6,57o 6,050 4,818 Average 3 5,776 6,407 6,213 5,184 JACKSON. 3,000 plants. 4,200 plants. 5,400 plants. Albemarle Prolific 8 .6,320 6,515 5.969 4,i34 Cocke Prolific 5 5,770 5,633 6,324 3,549 Hickory King 6 5,527 6,668 6,595 5,289 Huffman 6 8,872 9,640 10,422 3,983 Lewis Prolific 7 6,719 6,904 6,974 4,70i Little Willis 4 5,791 6,152 6,306 4,234 Webb Improved Watson. . . . 5 4,884 6,236 6,112 4,203 Average 5-9 6,269 6,821 6,958 4,299 a Calculated from selected factor. In the bottom-land series are shown the yields of 8 varieties which were grown from 5 to 13 years each, with a general average of 7.8 I 8 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. years. The best average rate pf planting for grain production for the 8 varieties was 6,774 stalks per acre. The average yields at all three rates of planting, however, are very close together. At the 8,000 rate, the total yield is only 2 percent more than at the 6,000 rate. At the 10,000 rate, which is an increase of 48 percent over the number of plants required for maximum grain production, the total crop was increased by less than 5 percent over that from the 6,000 rate and the calculated increase over the best rate for grain is less than 2 percent. In the upland series with 4 varieties for a period of three years, the best average grain rate was 5,184 stalks per acre, or a little less than the medium, or 5,400 rate. The average yield of air-dry crop at the 5,400 rate is 6,407 pounds, which is 194 pounds more than was obtained at the highest rate of planting. In the Jackson series, which included 7 varieties for an average period of 5.9 years, the best grain rate averaged 4,299 plants per acre. The average yield at the nearest rate of planting, the medium or 4,200 rate, was 6,821 pounds per acre. At the 6,600 rate of planting, with an increase of nearly 30 percent in the number of plants, the increased yield was only 137 pounds per acre. The evidence as obtained in the three series agrees in that the planting which gives the largest grain yield lacks little of giving the largest total yield of ear corn and stover. In fact, the results are so close that the doubt is left whether increasing the number of stalks above that required for the maximum grain production will increase the crop as a whole. However, as there is no evidence of loss from a moderate increase in rate of planting, and as in common farm prac- tice a thicker rate would be expected to result in a more complete occupation of the ground by reducing the number of gaps, an appre- ciable increase in the number of stalks per acre over that estimated for best grain production may be advisable for a silage crop. The Spacing of Corn. After the proper number of plants per acre has been determined, the question arises as to the effect on the yield both of the distance between rows and of the arrangement of plants in the row. Are wide-spaced rows as good as narrow? Does corn planted in check rows yield as well as corn planted in drills? Considerable data bear- ing on these questions have been obtained by different experiment stations, and the general conclusion has been reached that minor changes in arrangement, such as widening or narrowing the space MOOERS I CORN PLANTING RATES. 19 between the rows and planting in check rows or in drills, does not affect the yield, provided the number of plants per acre remains the same. There must always be limits, however, beyond which there is danger of decreased yield. The data presented in Tables 10 and 11 are given as an aid to the solution of the problem under conditions similar to those found in Tennessee. THE NUMBER OF PLANTS PER HILL. The data from experiments made in triplicate each year for five years on a soil of good fertility at the Knoxville station farm are given in Table 10. In these experiments the number of plants per hill varied from 1 to 3. The number of plants per acre in each case was 8,000, a number which is a little high for the fertility of the land. Table 10. — Yields of corn when planted in the hill at different rates, but with the same number of plants per acre. a Year. Number of plants Width of check Yield per acre. per hill. row. Grain. Stover. Feet. Bushels. Tons. 1907 2.3 x 2.3 50.6 2-34 2.3 x 2.3 42.6 1.58 1909 2.3 x 2.3 36.4 1.40 2.3 x 2.3 47.2 1.52 2.3 x 2.3 71.3 2.10 49.6 1.79 1907 2 j 3.3X3.3 51.8 2.27 1908 2 3-3 x 3.3 45-6 1.77 1909 2 3-3 X3.3 34-7 1-39 2 3-3 x 3-3 47.2 1.44 I9II 2 3-3 x 3.3 70.2 1.94 49.9 1.76 1907 3 4x4 48.4 2.15 I908 3 4x4 40.2 1.60 1909 3 4x4 38.0 1-33 I9IO 3 4x4 45-o 1.40 3 4x4 65-4 1.82 47-4 1.66 a Experiments conducted in triplicate each year on fiftieth-acre plots. The 5-year average yields are almost identical for the 1- and 2-plant stands. With the 3-plant stands, however, the yields average 2.2 bushels per acre less than with the 1 -plant, and 2.5 bushels less than with the 2-plant. The difference is noteworthy for the reason that in four out of the five years the lowest yields of grain were obtained 20 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. from the 3-plant stands and that without exception the lowest pro- duction of stover was from this planting. DISTANCE BETWEEN ROWS. Table 11 gives the yields of corn when planted in 7-foot and in 3.5-foot rows, but with the same number of plants per acre in each case. The results of the two years' trials agree in that the yields of both grain and stover are materially less from the 7-foot than from the 3.5-foot rows. In the 1914 trials the average yield per acre from the 7-foot rows was 75 bushels of grain and 3.42 tons of stover. The 3.5-foot rows produced an average of 88.4 bushels of grain and 4.81 tons of stover, or an increase of 13.4 bushels of grain and 1.39 tons of stover. In 191 5 both the upland and the rich alluvial soil gave similar results to those obtained in 1914, the average yield of the upland and alluvial soil experiments being 69.7 bushels of grain and 2.72 tons of stover from the 7-foot rows and 83.3 bushels of grain and 3.05 tons of stover from the 3.5-foot rows. Table ii. — Yields of corn when planted in 7-foot and in 3.5-foot rows. Year. Variety. Kind of soil. Plot No. Plants per acre. Yields per acre. 7-ft. rows. Grain. Stover. 3.5-ft. rows. Grain. Stover. 1914 1915 Albemarle do do Hickory King do do do do do Hickory King do Albemarle do do do do do Rich alluvial do do do do do do do do Upland do do do Rich alluvial do do do 8,000 8,000 8,000 7,000 7,000 7,000 7,000 7,000 7,000 Average 6,000 6,000 6,doo 6,000 8,000 8,000 8,000 8,000 Bushels. 92.1 90.0 65-3 67.0 67.6 67.9 Tons. Bushels. Tons. 4-30 3-77 j ! j 101.7 3-45 3-24 j 80.9 2.44 3-33 75-o 3-42 88.4 4.81 60.0 i-95 53-8 2-73 5i-4 2.85 44-8 1.49 113. 4.09 90.9 3-57 108.9 329 89.2 3-io Average . 69.7 I 2.72 82.6 4.66 4.80 4.98 83-3 3-05 MOOERS : CORN PLANTING RATES. 2 I GENERAL CONCLUSIONS FROM THE SPACING EXPERIMENTS. The two series of experiments, one to determine the best number of plants per hill for check-row planting and the other to determine the effect of material differences in distance between rows, supple- ment each other. The former series showed that no decrease in yield came from the narrow rows, which spaced the plants at equal dis- tances from each other. On the other hand, the latter series showed that the spaces between rows may easily be so wide as to reduce the yield. The general conclusion seems warranted, therefore, that the best results in practice will probably be obtained with a width of row which permits the satisfactory use of tillage implements but allows the determined number of plants to be as widely spaced as possible. Summary. 1. The rate of planting corn is of much practical importance in the South, but a definite rule is needed. Too thick planting is of com- mon occurrence. 2. Different varieties require appreciably different rates of plant- ing. In general the small and short-season varieties require thicker planting than the large, long-season varieties. 3. The experimental results indicate a close relationship between the best rate of planting for grain production and a definite yield of grain per plant. For example, the Hickory King variety was found to yield best when the rate of planting allowed the average plant to produce about 0.45 pound of grain. 4. To approximate the proper stand of corn a simple equation may be used, as follows : In this equation, N stands for the number of stalks per acre, Y for the expectancy, or approximate production in bushels per acre of the field in question under average seasonal conditions, and F is the standard varietal factor, or the average weight of grain per plant at the best rate of planting, as determined experimentally for the variety in question. 5. On land of high productiveness the average weight of grain per plant at the best rate of planting was found to vary from year to year with the nature of the season, so that the more favorable the season the greater was the weight. The standard varietal factor is practically the result obtained for the average of all seasons. 22 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. 6. The data as obtained may be interpreted as indicating that a varietal factor is practically a constant between the 36-and 65-bushel limits of productivity. However, the weight of evidence indicates that as the productiveness of the soil changes a change should be made in the varietal factor. On the average for 15 varieties, this change was calculated to be 0.0014 per bushel change in expected yield. With this in view, the following formulas may be used in con- nection with the table of standard factors : (1) For land producing on the average more than 64 bushels per acre — N 56V F+ (Y - 64) .0014" (2) For land producing on the average less than 64 bushels per acre — 56 Y N = F - (64 - F).ooi4 In addition the suggestion is made that no rate of planting should be less than that calculated for a 25-bushel crop. 7. The margin of safety appears to go with a too high rather than a too low rate of planting. 8. The results of three series of experiments agree in showing that the best rate of planting silage corn was little different from that which gave the highest yield of grain. In practice, however, a ma- terial increase seems warranted. 9. In the spacing experiments the conclusion was reached that the best results in practice will probably be obtained with a width of row which permits the satisfactory use of tillage implements, but allows the determined number of stalks to be as widely spaced as possible. PARKER: RUST RESISTANCE OF OATS. 23 A PRELIMINARY STUDY OF THE INHERITANCE OF RUST RESISTANCE IN OATS. 1 John H. Parker. Introduction. Having found that there are disease-resistant species or varieties in a given crop, the next problem often is one of hybridization. The literature of the subject contains many references to the discovery or production of disease-resistant varieties, but there are, in propor- tion, very few examples of the profitable or extensive culture of such varieties. This is largely due to the fact that most of them are so undesirable with regard to one or more other equally important char- acters that they cannot be generally and profitably grown. This is exactly the present condition of affairs in the case of rust-resistant oat varieties. It has been shown (15) 2 that so far as cultivated varieties are con- cerned, resistance to crown rust exists almost exclusively in varieties of the red oat group (Avena sterilis). Resistance to stem rust is not to be found in any of the varieties of the red oat group so far tested, but does exist in some of the varieties of the common oat group (Avena sativa). Some of the red oats resistant to crown rust are well suited to culture in the Southern States where they are widely grown, but they are not adapted to the conditions of the Northern States. On the other hand, the white oats, certain varieties of which are resistant to stem rust, seem specially adapted to the northern half of the United States and are entirely unsuited to condi- tions in the South. These conditions make it highly desirable that some knowledge be gained of the manner of inheritance of the char- acter, rust resistance. In fact, this is essential if the method of hybridization is to be intelligently and successfully used in producing 1 Contribution from the Office of Cereal Investigations, Bureau of Plant Industry, United State Department of Agriculture, Washington, D. C. Re- ceived for publication September 15, 1919. The experiments here described were conducted at Cornell University. The work was under the supervision of Dr. H. H. Love, Professor of Plant Breed- ing, whose suggestions and assistance are gratefully acknowledged. Thanks are due Mr. W. I. Fisher for making the photographs used in the paper. 2 Reference is made by numbers in parentheses to " Literature cited," page 37. 24 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. new varieties which are resistant to a given rust species, and are, in addition, adapted to local conditions and capable of producing a high yield of grain which will come up to market standards. The preliminary work reported in the present paper was done in ' the attempt to determine the genetic behavior of the character, rust resistance, in a cross between resistant and susceptible oat varieties of the two groups mentioned above. Historical Review. No attempt will be made to review completely the literature of dis- ease resistance, or even that of rust resistance in cereals. Orton (12, 13, 14) has succeeded in transferring the character, disease resistance, and has been able to obtain and fix races of hybrid progeny which combine this character inherited from one parent with others of com- mercial importance from the other parent. Norton (11), working on the control of asparagus rust, has achieved success thru using the method of hybridization, and the new rust-resistant race now is being- grown commercially. Farrer (6, 7) has succeeded in producing wheat hybrids which are rust resistant, and at least one of them is now commonly grown in Australia. Evans (5) has recorded rather dis- couraging results, for he has found what seems to be a case of a hybrid between a resistant and a susceptible wheat acting as a bridg- ing species. He was able, by culturing the rust on plants of the susceptible hybrid, to obtain spores which would infect the hereto- fore resistant parent, and which also increased the virulence of the infection on the susceptible parent. If these results should prove to be generally true, breeding for rust resistance in cereals would be of doubtful value. The work of Stakman and Piemeisel (17), however, has not strengthened the idea that bridging hosts are of great im- portance. Stakman, Parker, and Piemeisel (18) have shown that susceptible hybrids of the F x and F 2 generations of crosses between resistant and susceptible wheats did not enable the rust cultured on them to cause normal infection on seedlings of the resistant parent, nor to infect the susceptible parent more virulently. Biffen ( 1 ) crossed a wheat he classed as immune with a susceptible wheat and found that the Fj plants were susceptible and that in the F 2 generation the proportions of badly to slightly attacked were in the ratio of 3 to 1. In the F 3 generation the relatively immune individuals bred true to this character, but the susceptible types taken as a class produced either all susceptible offspring or a mixture of susceptible and relatively immune plants. PARKER: RUST RESISTANCE OF OATS. 25 In a second paper, BifYen (2) reports the results of further work and states that in counts made of F 2 plants from crosses between wheats immune from, and susceptible to, yellow rust, the ratio of 1 immune to 3 susceptible was again obtained. The F 3 generations of all these crosses were not studied, but in the F 2 there were indications of the existence of intermediate types. In the case of one cross between Rivet (immune) X Red King (susceptible) the extracted recessives only have been grown. These have maintained their immunity through the F 4 generation. The experiments with black stem rust are reported to have been " not altogether successful," but from analogy with the cases already described and from the susceptibility of the F 1 generation, it would appear that the susceptibility to black rust is also a dominant character. In a third paper, BifTen (3) has published an account of the con- tinuation of these earlier studies. In this paper he reports that the plant's of hybrid generations showed very varied degrees of susceptibility. Some were relatively slightly attacked, and between these extremes all degrees of susceptibility appeared to exist. In the work since 1907, where F 3 results were compared with the F 2 , is has been found : 1. That the immune types of the F 2 generation breed true to that feature. 2. That some of the susceptible forms also breed true but others produce off- spring showing segregation into susceptible and immune forms. 3. That plants of the " medium class " are not necessarily heterozygous. 4. Where segregation occurs, the sum total of immune and susceptible indi- viduals is again in the ratio of 1:3. The actual numbers of the whole series were 276 and 849. The results are those that might have been anticipated from those of the F 2 generation. They show, conclusively for this particular cross, that the varying degrees of susceptibility are not due to the effects of more than one " dose " of the factor concerned with the production of susceptibility. Having rejected the multiple-factor hypothesis as an explanation for the wide range in susceptibility observed, BifYen gives it as his belief that the varying and probably constant degrees of susceptibility seen in F 2 and succeeding generation probably are due to the extreme ease with which the degree of susceptibility is altered by slight changes in the plant's metabolism. Biffen further asserts that " none of the cases examined up to the present time indicates that rust susceptibility itself is due to the existence of more than one factor." Newman (9), in reporting on the work in cereal breeding at Svalof, gives the following account of the behavior of the character, rust re- sistance, in crosses : For example, from the crossing 0401 X 0705, both classified as grade 2 (mod- erately resistant), eight separate cultures were investigated during the bad rust year of 1904, and the differences observed in these plats were recorded as 26 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. exceedingly striking. The difference between the most susceptible culture and that which showed greatest resistance was much greater than the difference be- tween the parent's themselves. From the crossing 0319 X 0501 (both sorts of high resistance) there were produced a number of lines, some of which in F 3 proved much more susceptible than either parent. Thus it may be seen that apparently new forms may arise from crossing between sorts which to all appearances are practically identical in regard to certain characters. The origin of these forms is due simply to peculiar group- ing of definite units already in existence, and not to the acquisition of anything actually new. In other words, they constitute different gradations, a quanti- tative hereditary variation. The conclusions (English translation) reached by Nilsson-Ehle (10) in his study of the resistance of wheats to yellow rust (Puccinia glumarum) at Svalof , Sweden, are as follows : 1. Segregation of rust resistance in crossing is always definitely present. 2. The segregations in these investigations have always been complex. There customarily result from the segregation new gradations of rust resistance approaching the parental gradations, and in the F 3 plats, transgressions more susceptible or more resistant than the parents, are readly shown to exist. Also, crosses between varieties of about the same or differing only slightly in rust resistance give, in their progenies, segregation with trans- gressive forms. 3. The complicated segregation and the production of segregates from the crossing of lines of about the same resistance is explained by the existence of several independent (multiple) Mendelian factors, which influence (de- termine) the rust. All data found to date agree with this hypothesis. 4. The heritable gradations of rust resistance exhibited by the different varie- ties and lines are not due to independently arisen variations, but to differ- ent combinations of a number of factors, whereby different combinations of factors may exhibit about the same outer resistance. Through different new combinations of factors, new gradations arise, after the crossing of varieties or lines differing or of nearly the same rust resistance. Love (8) has called attention to the possibility, and perhaps the necessity, of using hybridization to obtain varieties of oats which shall be rust resistant, of good market quality, and of high yielding power. Description of Varieties and of Material Used. In the experimental studies here reported, pure lines of Burt and Sixty-Day oats and F 2 generation hybrid plants from a cross between these varieties comprised the material used. 3 The original cross was made by Doctor Love in the greenhouse at Cornell University during the winter of 1913-1914, and the F x generation plants were grown there by him the following winter, in connection with other studies of inheritance in oats. 3 From the cultures of the Department of Plant Breeding at Cornell University. PARKER : RUST RESISTANCE OF OATS. 27 The two parent varieties, Burt and Sixty-Day, are quite distinct in appearance, habit of growth, and particularly in the characters of the kernels. They represent different botanical groups, Burt usually being classified with the varieties derived from Avena sterilis and Sixty-Day being a variety of the Avena sativa group. Their origin, corresponding with that of the two botanical species, no doubt is dif- ferent, and their present distribution consequently is not the same. Both are early spring varieties, and the districts to which they are adapted overlap to some extent. Thus in Kansas and Nebraska both varieties are successfully grown. In the southern and southeastern States, Burt is probably the best spring variety to grow. Etheridge (4) gives a full description of Burt and Sixty-Day varieties. Warburton (19) gives brief notes on Burt, Sixty-Day, 4 and Kher- son oats, relating especially to their value as rust-escaping sorts. Plumb (16) states that Burt is rather locally grown, not being known over a wide territory. It does well, generally, in Mississippi, Alabama, Georgia, South Carolina, and northern Florida. It is grown to some extent in northern Alabama, Georgia, South Carolina, and middle and eastern Tennessee. The Burt oat is very similar to if not identical with Early Ripe and both were found to show resistance to crown rust. Because of this similarity in appearance and in reaction toward rust, an attempt was made to get some of the facts relating to the origin and history of the two varieties. Prof. C. A. Zavitz, of the Ontario Agricultural College, writes as follows regarding Burt and Early Ripe oats : 5 We obtained the Early Ripe variety from A. W. Livingston, seedsman, of Columbus, Ohio, in 1899, and the Burt from Dr. E. M. Freeman, of the U. S. Department of Agriculture, in 1906. We do not consider these valuable varieties, from an agricultural standpoint. In regard to rust, the Burt had 8 percent and the Early Ripe 6 percent, under similar conditions. There was only one variety, viz., Yellow Kherson, which has had less rust than the Early Ripe in the average experiment for 5 years. The Livingston Seed Company of Columbus, Ohio, writes as fol- lows, regarding Burt and Early Ripe oats : 6 We discontinued the handling of Early Ripe oats a number of years ago, and have none in stock at present, nor do we know of any in this section of the country. 4 In March, 1901, the United States Department of Agriculture received a variety under the name "Sixty-Day" from Mr. C. I. Mrozinski, of Proskurov, in the Podolia government of Russia. 5 Letter to the writer dated April 26, 1916. r ' Letter to the writer dated May 2, 1916. 28 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. We secured our stock when we commenced growing it of A. L. Reese, Shelby- ville, Ky. Where Mr. Reese secured the oats we do not know. It is our judgment that the oats that we listed at that time under the name of Early Ripe is identically the same as the oats which came out some years later under the name of Burt. However, we are not sure that they are exactly the same. Mr. J. H. Burrow, of Linnville, Tenn., in a circular letter, claims that this oat originated in Green County in south Alabama, in 1878, by a man by the name of Burt. Further than this we are unable to trace it. Whether Mr. Burrow is still living or not, we do not know, as we have not heard from him direct for a number of years. If there is any difference between Early Ripe and Burt oats, we were unable to distinguish what the difference was when growing them. We discarded them because there was not enough demand for them to justify continuing to keep them up from year to year. Mr. C. W. Warburton, of the U. S. Department of Agriculture, writes as follows regarding the Burt oat : 7 I do not know very much about the origin of the Burt oat, though it appar- ently was first grown in Tennessee. This variety is most extensively grown in Tennessee and adjoining states, from spring seeding. It also has done very well in our tests at Amarillo, Tex., and at Akron, Colo., and in varietal trials at the Nebraska station. It is strictly a southern oat, being of little value anywhere in the North. It is grown most extensively by farmers in Tennessee. In the northern half of the Southern States it is apparently one of the best varieties for spring sowing. It is not adapted to fall sowing, as it is not winter hardy. Prof. C. A. Mooers, of the Tennessee Agricultural Experiment Station, has written to Mr. Warburton as follows regarding the Burt : 8 I am sorry to say that I can not give you with any certainty the origin of our Burt oats. As you say, it has been grown in the State for a good many years. Some fifteen years ago I met an elderly gentleman, the editor of a local paper in middle Tennessee, who told me he knew the history of the Burt oats ; that it originated in Alabama, where a farmer noticed in his field of spring oats, presumably Red Rustproof, a plant which was ripe, while all the others around it were still green. He saved the seed, which was the beginning of the Burt oat. As I remember it, this was said to have occurred in the 6o's or 70's. Experimental Methods. The methods of sowing, inoculating, and note taking were the same as those described in detail in another paper (15), and hence a brief discussion of them will suffice here. The crown or leaf rust, Puccinia lolii avenae McAlpine, and the stem rust of oats, Puccinia graminis avenae Erikss. and Henn., were used in these experiments. Urediniospore cultures of both rusts were 7 Letter to the writer dated May 3, 1916. 8 Letter dated April 11, 1916. PARKER : RUST RESISTANCE OF OATS. 29 obtained from the Minnesota Agricultural Experiment Station and were increased on seedling plants of a susceptible oat variety, for use in inoculating the parent varieties and the hybrids. For the inoculations made on seedling plants, the seed was sown in* 4-inch pots, at the rate of 15 seeds per pot. The seedlings were later thinned to 8 or 10 per pot, the usual number inoculated. Sowings of the parent varieties and of the F 2 generation hybrid seed were made each week in series of about 10 or 12 pots. Thus, during the winter, there were always some sowings just coming up, other series ready for inoculation, and still others on which the uredinia had appeared. For the inoculations made on plants at time of heading, sowings were made during November in 5-inch pots. About 6 seeds per pot were sown, and the plants were later thinned to 2 or 3, which was the number usually inoculated and allowed to mature. The seedlings were inoculated when in the second leaf, and were then about 3 to 5 inches in height. All inoculations were made on the first seedling leaf, by scraping urediniospores from a leaf bearing mature uredinia, and applying them to the moistened leaf surface with a flattened steel needle. One pot of seedlings of Sixty-Day was inoculated with each series of hybrids, to serve as a control culture. Inoculations on the plants at time of heading were made in a similar manner, using crown rust on the upper leaf blades and stem rust on the upper sheaths and the peduncles. Notes were taken in from 6 to 10 days on the appearance of flecks and uredinia, and records were kept of the number of leaves inocu- lated, the number on which flecks or uredinia appeared, and of any other points necessary in determining the classification of the plants as resistant, intermediate, or susceptible. A plant was recorded as resistant when few or no normal uredinia ruptured the epidermis and where unmistakable signs of resistance were present. If the uredinia were numerous, large, and normal in appearance, and no very marked signs of resistance were present, the plant was included in the susceptible class. If there was a limited number of uredinia nearly normal in appearance, but accompanied by flecks, and if there were other indications of semi-resistance, the plant was classified as intermediate. A great majority of the inoculations were made on seedlings, for the reason that greenhouse space for growing the plants to maturity was limited* as was also the quantity of hybrid seed available. The amount of space available in the special moist chamber used for the older plants had also to be considered. 30 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. No records were available of the behavior toward rust of the two strains of Burt and Sixty-Day used in making the cross, when grown under field conditions at Ithaca. Neither is it known whether the F 1 generation plants were resistant or susceptible. The plants of the F 2 generation inoculated in the seedling stage could not be grown to maturity, and there are no results therefore beyond the F 2 generation. The results to be described, then, include only those with pure-line material of the Burt and Sixty-Day parents and of the hybrid plants of the F 2 generation. These belonged to five distinct families, or progenies of F l generation plants. Results of Experiments, stem-rust inoculations on seedling plants. The experiments with stem rust were not as extensive as those with crown rust, for the reason that neither the Sixty-Day, the Burt, nor any of the hybrids showed any resistance to stem rust. Only enough seedlings were inoculated to definitely prove this fact. There were 32 plants of Sixty-Day inoculated and all were found to be very susceptible. Fifty-two plants of Burt were used and gave similar results ( see Plate 1 , C) . Eighty-five F 2 generation hybrids were inocu- lated as seedlings and all showed complete susceptibility to stem rust. CROWN-RUST INOCULATIONS ON SEEDLING PLANTS. The experiments with crown rust gave quite different results, for altho one of the parent varieties, Sixty-Day, also proved to be com- pletely susceptible to this rust (see Plate 1, B and F), the other, Burt, gave definite evidence of being resistant. This was not true of all the seedling plants, for some were just as susceptible as those of Sixty-Day. Some of the hybrid plants also were extremely resistant, while others were entirely susceptible. The results of inoculations of crown rust on seedling plants of Burt oat are given in Table 1. ■ Table i. — Inoculations of crown rust on seedlings of Burt oat. Number. Percent. Seedlings classified as resistant 48 21.5 Seedlings classified as intermediate 152 68.2 Seedlings classified as susceptible 23 10.3 Total 223 100.0 These figures show that, altho the variety may be said to possess a rather high degree of resistance, there are many individual plants which are only semi-resistant and some which are quite susceptible. This fact also is shown very clearly in Plate 1. PARKER: RUST RESISTANCE OF OATS. 31 The fact that all of the Burt plants did not prove resistant shows that the strain used was not homozygous for the character, resistance to crown rust. This does not contradict the statement that the par- ental varieties were pure lines, for it is well known that a plant or line may be homozygous for some characters and heterozygous for others. This strain had, in fact, been found to be pure with regard to the usually observed agronomic characters. No satisfactory explanation of the occurrence of the semi-resistant and susceptible plants can be given, without knowing the hereditary behavior of the character in other generations. It does not seem likely that factors of environment or of the metabolism of the host plant can be given much consideration, for there often were marked differences in the degree of resistance manifested by plants grown in the same pot and inoculated at the same time from the same lot of urediniospores*. An effort was made always to keep all external factors as uniform as possible. The different degrees of resistance can hardly be considered as mere fluctuations, but should rather be attributed to differences in genetic constitution of the plants. Whatever may be the underlying cause of this sharp variation in resistance of the Burt plants, true resistance certainly is present in many of them and the results with the hybrids indicate that the par- ticular plants used in making the cross carried the factor or factors for resistance, for the resistant character appears unmistakably among plants of the F 2 generation. The most resistant segregates appear to be fully as resistant as any of the resistant individuals among the Burt parent. The lack of uniformity in the resistance of plants of the Burt strain, however, makes it more difficult to interpret the results obtained in the F 2 generation, for the genotypic constitu- tion of this parent as regards rust resistance is not known. The inoculations of crown rust on seedlings of the F 2 generation included 468 individuals. The results are shown in Table 2. Table 2. — Inoculations of crown rust on seedling of Burt X Sixty-Day hybrids. Family. No. of plants in- oculated. Resistant. Intermediate. Susceptible. Number. Percent. Number. Percent. Number. Percent. 2501 A-J. . 2501 A-P . 2501 A-R . 2501 A-H . 2501 A-N . 172 114 51 64 67 50 23 1 4 3 29.I 20.1 6.2 4.4 24 17 7 3 10 13.9 14.9 13.7 4.6 14.9 98 74 43 57 54 57-o 65.O 84.4 89.2 80.7 Totals . . 468 81 17-3 61 13.0 326 69.7 32 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. The basis of classification was the same as that used in the experi- ments with Burt, and has been described. It should be said that the word intermediate is not used in its strict genetic sense, for it is not known whether or not these semi-resistant plants will produce prog- enies resembling those of true hybrid intermediates (heterozygotes). The word refers rather to the relative quantity of rust, tho number of uredinia alone was never used as a determining factor in classify- ing a plant. Plants put in the intermediate class usually showed some signs of resistance, such as unusually small uredinia, presence of flecks, etc., but had too many nearly normal uredinia to be classed as resistant. The classification is an arbitrary one and is based only on phenotypic appearance, tho it is probable that the plants in the resist- ant group belong to a common genotype. True segregation occurred and was plainly evident in the hybrid plants. This is indicated by the numerical results and is clearly shown in Plate 2. On the very resistant plants almost no uredinia were able to rupture the epidermis, while flecks and larger areas of dead host tissue, so characteristic of resistant varieties, are plainly visible. On the susceptible plants there are normal uredinia, the surrounding leaf tissues are still green, and there are no indications of resistance. The segregation was just as sharp as that observed in crosses in- volving morphologic characters, while some of the hybrid groups showed what appeared to be almost a complete series of transition forms between the wholly susceptible and the very resistant extremes, as illustratd in Plate 2, C. Such a series often is obtained in the F 2 generation where size characters or other complex quantitative differ- ences are involved. Altho all the plants were put into three groups or classes, it was clearly recognized that all of those in the intermedi- ate class were not exactly alike phenotypically, nor is it to be assumed that they all belong to the same genotype. In fact, they probably rep- resent a rather complete series, which closely approaches the suscept- ible group at one extreme and the resistant group at the other. It would be very difficult, if not impossible, to measure accurately and rank properly the individuals of such a series. There is no definite standard which can be used, as there is for size differences which can be measured and recorded in figures, or those where the results of chemical analyses may be used. It is probable that most of the indi- viduals in the intermediate class would be described as resistant, if considered in direct comparison with those of a very susceptible variety. They did not, however, show the extreme resistance of some of the Burt plants. Journal of the American Society of Agronomy. Plate 1 Journal of the American Society of Agronomy. Plate 2. PARKER: RUST RESISTANCE OF OATS. 33 Whether the figures from each of the five hybrid families are con- sidered separately or whether merely the totals are used, they show very clearly and positively that there were many more susceptible than resistant individuals. The total figures show that if only the two groups, resistant and susceptible, are considered, there were ap- proximately four times as many susceptible as resistant individuals. These results suggest that rust resistance is the recessive character and rust susceptibility the dominant one, and agree in this respect with Biffen's work with wheat, mentioned above. It is evident from Table 2 that the percentage of resistant plants was considerably lower in the last three families than in the first two. These results may be explained on the assumption that the Burt par- ent plant was heterozygous as to rust resistance. The study of the Burt material itself strengthens this view and, if it is correct, then different F 2 generation families, i. e., progenies from distinct F t generation plants, would be expected to have different ratios of re- sistant and susceptible individuals. INOCULATIONS ON PLANTS AT TIME OF HEADING. Four culms of the Sixty-Day oat were inoculated with stem rust. Two were heavily rusted, the other two produced a few normal uredinia. Of the four upper leaf blades which were inoculated with crown rust, three were very heavily rusted, and the fourth had some normal uredinia. These upper leaf blades infected with crown rust are shown in Plate I, F. These results on plants inoculated at time of heading confirm the fact established by the experiments with seed- lings, that the Sixty-Day variety is susceptible to both stem rust and crown rust. Six culms of the Burt oat were inoculated with stem rust. Heavy infections with large, confluent uredinia developed on three, while on the other three the uredinia were normal but fewer in number. Six upper leaf blades were inoculated with crown rust and, altho some normal uredinia were produced on all the leaves, none had a large number of them. A more significant fact was the occurrence of numerous flecks, a condition characteristic of resistant plants. The very sharply contrasted types of infection resulting from crown- rust inoculation on leaves of the two parent varieties, Burt and Sixty- Day, are shown in Plate I, E and F. The conclusions reached from the results of inoculations made at the time of heading on plants of the Burt parent are the same as those arrived at from the experi- ments with seedlings, i. e., that the Burt strain used is completely 34 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. susceptible to stem rust, but that at least some plants are resistant to crown rust. Eight culms of the F 2 generation hybrid plants were inoculated with stem rust, and heavy infections resulted on all of them. Eight upper leaves were inoculated with crown rust. The infection on one was very heavy; on one fairly heavy; on four slight, with no normal uredinia ; on one flecks only were evident ; and on one there were no visible signs that infection had taken place. The upper leaf blades shown in Plate 2, D, were selected from the series just de- scribed, and offer good evidence of the segregation of resistant and susceptible plants in the F 2 generation. The results from the limited number of hybrid plants which were inoculated at time of heading fully confirmed those obtained from seedlings, and show that the hybrid plants were all susceptible to stem rust, but that there were segregates which were resistant to crown rust and others which were susceptible. Discussion of Results. On the basis of this preliminary work, it is hardly possible to at- tempt the construction of a factorial hypothesis to fit the results. It is not desired, either, to put much emphasis on the ratios obtained, because of these facts : ( i ) The F t generation plants were not ob- served, and nothing is known of their resistance or susceptibility; (2) the number of F 2 generation plants observed was relatively small, and as these were not grown to maturity, the problem was not studied in the F 3 generation; (3) one of the parent varieties apparently was in a heterozygous condition for the character, rust resistance. The fact that in the F 2 generation more than three times as many susceptible as resistant segregates appeared indicates that suscepti- bility to crown rust is partially dominant and that resistance to the same rust is recessive. This question of dominance could be definitely answered only by knowing the behavior of the F 1 generation plants, when grown under conditions favorable for rust. The question of dominance and recessiveness is not one of major importance and little emphasis is placed on it by more recent investigators of genetic problems. The facts of segregation and recombination of characters according to Mendelian laws are of much greater significance. The question of linkage also is now recognized as being of great im- portance, because of its possible interference in the production of new and desired factor combinations. There can be no doubt that segregation did occur with respect to PARKER : RUST RESISTANCE OF OATS. 35 the character, resistance to crown rust. Little is known, however, as to the particular recombinations which took place, that is, of the rela- tion of rust resistance to other inherited characters. Neither is the real nature of the plants classified as intermediates known. Only by means of further breeding experiments can it be shown whether or not they are true hybrid intermediates, and what are their genotypic or factorial constitutions. Until the basis of rust resistance is much better understood than at present, it may not be possible to discover the exact manner in which the character is inherited. The most important fact brought out in the course of these experiments is that resistance to crown rust is a heritable character and that it appears in a certain number of the F 2 generation plants. Other plants of the same generation show varying degrees of resistance, ranging all the way to complete sus- ceptibility. It is not well to go far into generalizations, but it may be said that as these results have been obtained in the cross, Burt X Sixty-Day, it is likely that similar ones would follow in other crosses between varie- ties of the Avena sterilis group which are resistant to crown rust and varieties of the Avena sativa group which are susceptible. Contrasted characters such as resistance and susceptibility almost certainly involve extremely complex physiological processes, perhaps to an extent modified by outside factors but primarily determined by heredity. The F 2 generation results, especially the difficulty of classi- fication, indicating the presence of a rather complete series of forms, do not favor the view that the character is a simple one, determined by a single factor difference. It seems more likely, as Nilsson-Ehle has found with regard to yellow rust (Puccinia glumarum) in wheat, that the observed gradations between resistant and susceptible plants are due to several determiners recombined by crossing. In other words, the character of rust resistance is probably dependent on mul- tiple factors. At least, that hypothesis will account for the results so far obtained, altho the number of these factors, their basis, and their exact behavior in heredity are still to be determined. Summary. i. Pedigree lines of two oat varieties, Burt and Sixty-Day, to- gether with a large number of F 2 generation hybrids between these varieties, were studied in relation to their rust resistance. Most of the inoculations were made on seedlings, but enough were made on plants at time of heading to show that the results were similar. 36 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. 2. The rusts used were the crown rust of oats, Puccinia lolii avenae McAlpine, and the stem rust of oats, Puccinia graminis avenae Erikss. and Henn. 3. Burt and Sixty-Day and all the hybrids of these two varieties so far tested were found to be entirely susceptible to stem rust. 4. All plants of Sixty-Day also were uniformly susceptible to crown rust. Of 223 inoculated plants of Burt, 48 were classified as resist- ant, 152 as intermediate, and 23 as susceptible. 5. Each of the five hybrid families contained, in the F 2 generation, some plants showing a high degree of resistance to crown rust and others which were as susceptible as plants of the Sixty-Day parent. In other words, there was definite segregation. There was, however, a rather large number of plants which were classified as intermediate and which showed varying degrees of resistance. 6. The numerical results of inoculations made in the F 2 hybrids were as follows : <• Number. Percent. Seedlings classified as resistant 81 17.3 Seedlings classified as intermediate 61 13.0 Seedlings classified as susceptible 326 69.7 Total 468 100.0 7. The fact that there were so many more susceptible than resistant plants indicates that susceptibility to crown rust in this cross is par- tially dominant, while resistance is recessive. 8. These contrasted characters are not thought to be due to en- vironmental conditions or to differences in the metabolism of the host plants, but to definite genetic factors. Nonhereditary factors may of course influence or modify their expression. 9. Rust resistance and susceptibility hardly can be considered as simple characters or as being determined by a single factor difference. 10. The F 2 generation results, particularly the rather complete series of forms showing varying degrees of resistance and necessarily classified as intermediates, favor the view that several factors are involved, i. e., the multiple factor hypothesis. 11. No attempt has been made to construct a definite factorial hy- pothesis to explain the results obtained. 12. This preliminary work has proved the possibility of transfer- ring the character of resistance to crown rust from the Burt variety to plants of the F 2 generation of a Burt X Sixty-Day cross. 13. This suggests further use of the method of hybridization in the effort to produce rust-resistant varieties of oats for culture in the several oat districts of the United States. PARKER : RUST RESISTANCE OF OATS. 37 Literature Cited. 1. Biffen, R. H. Mendel's law of inheritance and wheat breeding. In Jour. Agr. Sci., v. I, pt. I, p. 4-48, pi. 1. 1905. 2. . Studies in the inheritance of disease resistance. In Jour. Agr. Sci., v. 2, pt. 2, p. 109-128. 1907. 3. . Studies in the inheritance of disease resistance II. In Jour. Agr. Sci., v. 4, pt. 4, p. 421-429. 1912. 4. Etheridge, W. C. A classification of the varieties of cultivated oats. N. Y. (Cornell) Agr. Expt. Sta. Memoir 10. 1916. 5. Evans, I. B. P. South African cereal rusts, with observations on the prob- lem of breeding rust-resistant wheats. In Jour. Agr. Sci., v. 4, pt. 1, p. 95-104. 191 1. 6. Farrer, William. The making and improvement of wheats for Australian conditions. In Agr. Gaz. N. S. Wales, v. 9, pt. 2, p. 131-168; pt. 3, p. 241-260, 1 pi. 1898. 7. . Report of the wheat experimentalist. In Agr. Gaz. N. S. Wales, v. 15, pt. 11, p. 1047-1050. 1904. 8. Love, H. H. Methods of breeding oats. Cornell Reading Courses, v. 2, no. 44, p. 185-200. 1913. 9. Newman, L. H. Plant breeding in Scandinavia, 193 p., illus. Ottawa, 1912. 10. Nilsson-Ehle, Herman. Kreuzungsuntersuchungen an Hafer und Weizen, II. In Lunds Univs. Arsskr., afd. 2, bd. 7, no. 6, 84 p. 191 1. 11. Norton, J. B. Methods used in breeding asparagus for rust resistance. U. S. Dept. Agr., Bur. Plant Indus. Bui. 263, 60 p., 4 fig., 18 pi. 1913. 12. Orton, W. A. On the theory and practice of breeding disease-resistant plants. In Proc. Amer. Breeders' Assoc., v. 4, p. 144-156, 7 fig. 1908. 13. . The development of farm crops resistant to disease. In U. S. Dept. Agr. Yearbook 1908, p. 453-464, pi- 39~40. 1909. 14. . The development of disease-resistant varieties of plants. In 4e Conf . Internat. Genetique, Compt. Rend. Acad. Sci. (Paris), 191 1, p. 241-265, 9 fig. 1913. [5. Parker, John H. Greenhouse experiments on the rust resistance of oat varieties. U. S. Dept. Agr. Bui. 629. 1918. 16. Plumb, C. S. The geographic distribution of cereals in North America. U. S. Dept. Agr., Div. Biol. Survey Bui. 11, 24 p., 3 fig., 1 col. map. 1898. 17. Stakman, E. C, and Piemeisel, F. J. Biologic forms of Puccinia graminis on wild grasses and cereals. A preliminary report (Abstract). In Phytopath., v. 6, no. 1, p. 99-100. 1916. [8. Stakman, E. C, Parker, John H., and Piemeisel, F. J. Can biologic forms of stem rust on wheat change rapidly enough to interfere with breeding for rust resistance? In Jour. Agr. Research, v. 16, no. 2, p. 111-123. 1918. 19. Warburton, C. W. Sixty-Day and Kherson oats. U. S. Dept. Agr. Farm- ers' Bui. 395, 27 p., 5 fig. 1910. Legends for Plates. Plate i. A. Seedling leaves of Burt oat inoculated with crown rust, show- ing resistant (left) and susceptible (right) strains. B. Seedling leaves of Burt oat (left) resistant to crown rust, and of Sixty- Day oat (right) susceptible. 38 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. C. Seedling leaves of Burt oat inoculated with crown rust (left), showing sharp resistance, and inoculated with stem rust (right), showing susceptibility. D. Upper leaf blades and sheaths of the same plants as in Plate 2, E, show- ing susceptibility of all three to stem rust and susceptibility of Sixty-Day to crown rust, contrasted with sharp resistance to crown rust of the hybrid and Burt. E. Upper leaf blades of resistant Burt oat, from plants inoculated with crown rust at time of heading, showing flecks, indicative of resistance, but almost no uredinia. F. Upper leaf blades of susceptible Sixty-Day oat', from plants inoculated with crown rust at time of heading, showing normal uredinia. Plate 2. A. Seedling leaves of F 2 hybrids (Burt X Sixty-Day) inoculated with crown rust, showing resistant (left) and susceptible (right) segregates. B. Same as A, but of another F 2 family. C. Same as A, but of another F 2 family. D. Upper leaf blades of F 2 hybrid plants (Burst X Sixty-Day) inoculated with crown rust at time of heading. One leaf (left) has numerous uredinia, others have very few uredinia but show sharp flecks, indicating resistance. E. Plants of Sixty-Day, F 2 hybrid 1 , and Burt oats (left to right), inoculated at time of heading, showing crown rust on upper leaf blades and stem rust on upper leaf sheaths. richey: formaldehyde treatment of seed corn. 39 FORMALDEHYDE TREATMENT OF SEED CORN. 1 Frederick D. Richey. In an effort to prevent the growth of fungi on corn seedlings grown in water culture, the seed was first treated with formaldehyde solu- tions of different strengths and for different periods. This seed was then tested for germination and notes made on the early growth of the seedlings both in water culture and, for purposes of comparison, in sand. The results obtained from these tests form the basis for the present paper. Methods and Materials. The samples used consisted of 25 kernels, taken 1 kernel from each of 25 ears, of U. S. Selection 201 grown at Armorel, Ark., in 1918. Each treatment was given to 4 samples. The strength of the solutions is given as cubic centimeters of " Liquor Formaldehyde, U. S. P." (containing about 39 percent H-CHO) per liter of solution, ordi- nary tap water which had been boiled for 30 minutes and then cooled being used for dilution. The samples were soaked in the solutions for 30 minutes or for 2 hours, as indicated, after which the surplus was drained off and the containers closed for a period of fuming or penetration. The time of soaking plus the time of fuming gives the total period of treatment. Germination Experiments. Two sets of samples were tested for germination in water germi- nators designed by Mr. C. H. Kyle, in which the kernels were sup- ported on trays of woven wire cloth so that they could be immersed in water to the desired depth. For purposes of comparison, a set was also grown in sand ; while to ascertain the residual effect, one set of samples was dried to its original weight, held for 16 days, and then grown in sand. One untreated sample and three which had been soaked in boiled tap water were germinated in each case as controls. The experiments were conducted in a greenhouse in which the tem- perature usually approximated 8o° F. and varied from 70 to 87 F., which may be taken for the germination temperatures of the samples in water germinators. The temperature 1.5 inches below the surface 1 Contribution from the Bureau of Plant Industry, United States Department of Agriculture, Washington, D. C. Received for publication October 30, 1919. 40 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. of the sand (the depth at which the plantings were made) approxi- mated yo° F. thruout the experiment. Daily counts were made of the radicles and plumules appearing on the kernels in the water germinators, of the seedlings infected with fungi, and of those in which infection was severe. Germination in sand was based on the number of plants emerging. After germina- tion was apparently completed, the plants were measured and the rate of growth after emergence was calculated. RESULTS IN WATER CULTURE. Table I shows the percentage germination of the samples based on counts of radicles and of plumules, and the rapidity of germination and growth based on the emergence and measurements of plumules. Table i. — Germination and growth of corn in water culture as influenced by treatment with formaldehyde. Sam- Treatment. Seed producing radicles. Seed producing plumules. Average rapidity pie No. Formal- dehyde per liter. Time soaked. Total period. Box 1. Box 2. Aver- age. Box 1. Box 2. Aver- age. Germi- nation. Growth per day. 13 22 5 14 C.C. No Min. treatm 120 120 120 Hours. ent 3-5 8-5 26.5 Per- cent. 100 100 100 100 Per- cent. 96 100 100 100 Per- cent. 98 100 100 100 Per- cent. 72 76 84 60 Per- cent. 64 •84 80 72 Per- cent. 68 80 82 66 Days. 3-44 3-n 3-05 2.61 Inches. 1. 129 1.080 1.268 1. 041 2 23 19 5 5 5 30 30 30 3-5 8-5 26.5 100 100 100 100 96 96 100 98 98 100 100 100 96 96 88 98 98 94 3-59 3.02 3-43 1.432 1.329 1-374 3 10 6 5 5 s 120 120 120 3-5 8-5 26.5 100 100 100 96 100 100 98 100 100 96 96 96 96 80 100 96 88 98 3-44 3-30 2.98 1-352 1.226 1. 212 4 12 8 15 15 15 30 30 30 3-5 8-5 26.5 96 100 100 100 100 96 98 100 98 96 92 96 100 96 92 98 94 94 3-21 3-69 3-48 1.229 1-333 1. 115 24 20 15 15 15 15 120 120 120 3-5 8-5 26.5 96 100 96 100 100 96 98 100 96 96 92 80 100 100 80 98 96 80 3-30 3-38 3-38 1.230 1. 156 1. 103 25 21 16 25 25 25 30 30 30 3-5 8-5 26.5 100 92 88 100 92 72 100 92 80 96 80 80 100 84 56 98 82 68 3-56 3-42 3-7i 1-365 1. 151 0.852 11 7 17 25 25 25 120 120 120 3-5 8-5 26.5 100 84 76 96 88 36 98 86 56 80 72 56 84 68 24 82 70 40 3-66 3-83 3.62 1. 120 0.908 0.765 RICHEV : FORMALDEHYDE TREATMENT OF SEED CORN. 4I The samples treated with the 25 c.c. solution for 8.5 and 26.5 hours evidently were injured, as shown by the low percentage of radicles and still lower percentage of plumules. Samples 15 and 11, tho not showing any injury when measured by the percentage of radicles, were unable to develop their full number of plumules. The quantita- tive germination of the other treated samples does not seem to have been affected, the low germination of sample 10 in box 2 evidently being accidental. Altho the untreated and water-treated samples showed practically perfect germination, based on radicle counts, they were so severely infected by fungi that they were unable to put forth plumules, and show poor germination based on plumule counts. Table 2. — Influence of formaldehyde treatment on fungous infection of seed- lings in water culture. Treatment. Infected seedlings at stated periods from beginning of test. Sciinplc No. Formalde- hyde per 'liter. Time soaked. Total period. 3 days. 4 days. 5 days. 6 d Total. ays. Severe. C.C. Minutes. Hours. Percent. Percent. Percent. Percent. Percent. 13 No treatment 41 44 76 91 59 120 3-5 30 49 95 95 85 120 8-5 34 53 68 80 5i 120 26.5 67 76 88 97 73 2 5 30 3-5 5 11 13 27 4 23 5 30 8-5 4 25 47 16 19 5 30 26.5 4 21 45 21 3 5 120 3-5 10 19 6 10 5 120 8-5 9 23 2 6 5 120 26.5 16 37 2 4 15 30 3-5 2 10 3i 10 12 15 30 8-5 7 43 6 8 15 30 26.5 11 45 13 24 15 120 3-5 10 4i 12 20 15 120 8-5 9 44 15 15 15 120 26.5 3 11 63 23 25 25 30 3-5 4 27 10 21 25 30 8-5 10 37 5 16 25 30 26.5 3 47 6 11 25 120 3-5 3 8 49 5 7 25 120 8-5 11 29 3 17 25 120 26.5 O 70 30 The rapidity of germination and growth, while not materially af- fected by treatment with the 5 c.c. or 15 c.c. solutions, was checked 42 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. by the use of the 25 c.c. solution. The seedlings from the samples treated with the 5 c.c. solution were the most uniform, having an average coefficient of variability of 32.0 percent against 47.9 percent for the untreated and water-treated samples and 41.8 and 49.8 per- cent for those treated with the 15 and 25 c.c. solutions, respectively. Fungous infection was apparent on the sixth day on 46 of the 53 seeds which failed to germinate, and on 98 of the 112 which developed radicles but no plumules, while in both these classes the few unin- fected kernels were distributed among the samples without apparent relation to the treatment. The infection on the seedlings, which was quite severe in the case of the untreated samples, was markedly checked by all the treatments, being held practically to nothing until the fifth day, particularly in the case of samples 3, 10, and 6 (Table 2). RESULTS OF GERMINATION IN SAND. Table 3. — Germination and early growth of corn in sand as influenced by treatment with formaldehyde. Sample No. Treatment. Tested immediately after treatment. Tested 16 days after treatment. Formal- dehyde per liter. Time soaked. Total period. Germination. Growth per day. Germination. Growth per day. Percent. Rapidity. Percent. Rapidity. C.C. Minutes. Hours. Days. Inches. Days. Inches. 13 No treatment 96 5.58 1.387 IOO 4-33 1.031 22 120 3.5 96 5-92 1.338 96 4.44 0.883 5 120 8.5 96 4.96 1-397 IOO 4.22 0.926 14 120 26.5 IOO 4.96 1.429 IOO 4-25 1. 013 2 5 30 3-5 100 5.38 1-437 IOO 4-43 0.858 23 5 30 8-5 IOO 5-7i 1.203 IOO 4-56 0.807 19 5 30 26.5 96 5-58 1. 217 IOO 4.67 0.777 3 5 120 3-5 96 5-54 1-345 IOO 4.60 0.869 10 5 120 8-5 IOO 5-46 1.251 IOO 4.44 0.866 6 5 120 26.5 96 5-33 1.259 96 4-39 0.852 4 IS 30 3-5 IOO 5-42 1. 170 88 5.00 0.705 12 15 30 8-5 96 5-67 1.088 88 5-51 0.696 8 15 30 26.5 96 5-75 1.068 76 5-54 0.617 24 15 12 3-5 IOO 6.00 1. 150 40 5-55 0.514 20 15 120 8-5 96 5-79 1.040 60 5-92 0.526 IS 15 120 26.5 92 5-79 1.043 72 5-68 0.596 25 25 30 3-5 IOO 6.08 1. 112 40 5-75 0.634 21 25 30 8-5 IOO 5-96 1.030 24 6.21 0.486 16 25 30 26.5 60 6.29 0.863 16 7-25 0.400 11 25 120 3-5 96 6.00 1. 105 8 6.63 0.572 7 25 120 8-5 92 6.29 1.032 8 5-25 0-435 17 25 120 26.5 40 7.04 0.878 RICHEY : FORMALDEHYDE TREATMENT OF SEED CORN. 43 Table 3 gives the percentage of germination and the rapidity of germination and growth in sand, both as tested immediately after treatment and 16 days after treatment. The samples tested immedi- ately after treatment gave practically perfect germination except 16 and 17, in which there was evidence of injury. Of the samples held for later testing, the untreated sample, those treated with water only, and those treated with the 5 c.c. per liter solution gave perfect ger- mination, while the germination of the samples treated with the 15 c.c. and 25 c.c. solutions was reduced, the longer and stronger treat- ments resulting in the lowest percentages of germination. The rapidity of germination in either sand test was not materially affected by treatment with the 5 c.c. solution, while it was somewhat checked by the use of the 15 c.c. and 25 c.c solutions. The rapidity of growth was lessened by all treatments, though not materially so by treatment with the 5 c.c. solution. Conclusions. Treatment of seed corn with solutions of 5, 15, and 25 c.c. of for- maldehyde per liter materially reduced the development of fungi on the plants grown in water culture. The vitality of the seed, as evidenced by the development of the seedlings in either water culture or sand, was not affected by treat- ment with the solution of 5 c.c. of formaldehyde per liter. Treat- ment with the solution of 15 c.c. of formaldehyde per liter did not materially affect the seedlings grown in water culture, but neverthe- less was injurious as evidenced by the germination and development in sand. Treatment with 25 c.c. per liter was markedly deleterious. The germination and growth of samples 3, 10, and 6 was not af- fected by the treatments given, and the seedlings were remarkably free from fungi. Soaking seed corn for 2 hours in a solution of 5 c.c. Liquor Formaldehyde in 995 c.c. of water, followed by a fuming period of from 2 to 24 hours, can therefore be recommended as check- ing fungous development, without interfering with the normal devel- opment of corn seedlings in water culture. 2 2 Investigation of disease prevention in corn culture was no part of the pur- pose of the foregoing experiments, nor were they made with any specific fungus in mind. Microscopic examination of eight of the severely infected seedlings from the water culture, however, showed the typical crescent shaped macro- conidia of Fusarium to be present in every case. It is also known that the field in which the seed used in these tests was grown in 1918 was rather heavily infested with Fusarium. With these facts in mind, the marked reduction of fungous infection on the plants from the seed treated with formaldehyde indicates the possibility of reducing the amount of Fusarium in cultivated ; fields, insofar as this may be due to spores carried externally on the seed. 44 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. AGRONOMIC AFFAIRS. MEMBERSHIP CHANGES. The membership of the Society reported in the November, 1919, issue was 550. Since that time, 11 new members have been added, making a total membership of 561. The names and addresses of the new members, with such changes of address as have been reported, follow. New Members. Carlson, F. A., Dept. Soil Technology, Cornell University. Coffman, F. A., Akron Experiment Farm, Akron, Colo. Gordon, F. E., Hardin, Mont. Graves, Chas. L., State College, Brookings, S. Dak. Hart, J. C, Blacksburg, Va. Helmick, Ben C, 13 Pond St., Orono, Maine. Hendrix, W. R., Agronomy Dept., La. State Univ., Baton Rouge, La. McFadden, E. S., Highmore Substation Highmore, S. Dak. Machlis, Kos. A., Station A, Brookings, S. Dak. Swanson, A. F., Experiment Station Hays, Kans. Worthen, E. L., Dept. Soil Technology, Cornell Univ., Ithaca. Changes of Address. Trout, C. E., Oakhurst Farms, Jacksonville, Fla. Van Nuis, C. S., 134 Livingston Ave., New Brunswick, N. J. Wermelskirchen, Louis, Glenwood, Iowa. Wilkins, F. S., Iowa State College, Ames, Iowa. Wolfe, T. K., Box 413, Ithaca, N. Y. NOTES AND NEWS. P. V. Cardon, who has been superintendent of the Judith Basin Experiment Farm, Moccasin, Mont., for the past two years, has been elected professor of agronomy in the Montana State College and agronomist of the station, succeeding Alfred Atkinson, now president of the Montana college. Mr. Cardon will begin his new duties on March 1. J. Stanley Cobb has been appointed instructor in agronomy at the Pennsylvania State College, and J. Stanley Owens has been made in- structor in agronomy extension in the same institution. Carl G. Degen has been appointed assistant in agronomy at the Pennsylvania station. W. R. Dodson, dean of the college of agriculture of Louisiana State University and director of the Louisiana station for the past twenty-five years, has resigned, effective January 1. JOURNAL OF THE American Society of Agronomy Vol. 12. February, 1920. No. 2 FEDERAL SEED GRAIN LOANS. 1 C. W. Warburton. For the first time in the history of the United States, direct loans have been made by the Federal Government to farmers to enable them to meet conditions of financial stress caused by crop failures. On several previous occasians relief has been provided by Congress for those whose crops had been destroyed by floods or other calamity, but this relief has been in the form of the free distribution of seed without expectation of return. For the past three years the Federal Land Banks have made loans on farms, but their funds are derived from the sale of farm-mortgage bonds and the loans are secured by mortgages on improved farm land. In the present instance, Federal funds were loaned to farmers who could give no security other than a mortgage on the crop which they expected to produce. Because these loans are without precedent in our history, it seems worth while to record the conditions under which they were made and to discuss some of their agronomic phases. The demand for Federal aid to enable farmers to maintain or to increase their acreage in food crops came first from Kansas, where a very large part of the wheat sown in 1916 and 1917 had been de- stroyed by drouth and winterkilling. A delegation from that State came to Washington in June, 191 8, and presented figures to the Con- gressional committees on agriculture which indicated that, if Federal aid was not given, a greatly reduced acreage of wheat would be sown 1 Contribution from the Bureau of Plant Industry, United States Depart- ment of Agriculture, Washington, D. C. Read by title at the twelfth annual meeting of the American Society of Agronomy, Chicago, 111., November 10, 1919, in the absence of the author. 45 46 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. in Western Kansas in the fall of 191 8. According to the estimates presented, it was probable that not more than 1,300,000 acres of wheat could be sown in the western half of the State without Federal aid, whereas double this acreage had been sown the previous year and would be sown in 191 8 if aid was given. Similar reports came from western Oklahoma, and a little later requests for Government aid came from Montana, where drouth in 1917 and 1918 had greatly reduced the yields of wheat and other crops. Several bills were introduced in Congress during the summer of 1 91 8 appropriating various sums of money for extending aid to farm- ers in the drouth-stricken areas, but is soon became apparent that action on these could not be obtained quickly enough to help farmers put in their fall-sown crops. Late in June, 1918, Congress had author- ized an appropriation of $100,000,000 as a war emergency fund for "the national security and defense," which could be used by the President for any purpose that in his opinion would help to win the war. On July 26, 1918, on recommendation of the Secretary of the Treasury and the Secretary of Agriculture, the President set aside $5,000,000 of this fund for farmers' seed grain loans. The conditions of the loans were set forth in Joint Circular No. 1 of the Treasury Department and the Department of Agriculture, the President having designated these two departments to administer the fund. The Fed- eral Land Banks of the respective districts in which the drouth- stricken areas were located were named as financial agents of the United States to make and collect the loans. For purposes of administration by the Department of Agriculture, the drouth-stricken areas were divided into two districts, the northern one including western North Dakota and northern Montana, and the southern one including western Kansas, western Oklahoma, north- western Texas, and northeastern New Mexico. Mr. G. I. Christie, then Assistant to the Secretary, was designated to supervise the work in the northern district with headquarters at Great Falls, Mont., and Mr. L. M. Estabrook, Chief of the Bureau of Crop Estimates, that in the southern district, with headquarters at Wichita, Kans. Mr. H.N. Vinall and the writer assisted in the supervision of the work in the southern and northern districts, respectively. During August, several preliminary meetings were held in each district at which the condi- tions of the loans were explained by representatives of the Federal Land Banks and the Department of Agriculture, and plans were made for furthering the work. These meetings were attended by officials of the State extension departments, Farm Bureau presidents, county WARBURTON : FEDERAL SEED LOANS. 47 agents, and other interested persons. While considerable dissatis- faction with the terms of the loan was expressed by some of those attending these meetings, the general opinion was that the plan should be given a thoro trial and that the best possible use should be made of the funds. The agronomic features of the loan were under the general super- vision of the Department officials previously mentioned, assisted by the extension divisions of the various States, including the county agents in the affected counties. Fortunately, county agents were located in a large portion of these counties, and in many there were farm bureaus, so that the organization was ready at hand. The local banks were designated by the Federal Land Banks as agents to receive applications for loans, and on the completion of certain preliminaries, the proceeds of the loans were forwarded to the farmers thru these banks. In Joint Circular No. I, it was stated that: The primary object of farmers' seed grain loans is not to stimulate the plant- ing of an increased acreage of grain in the drouth areas, or even necessarily to secure the planting of a normal acreage, but rather to assist in tiding the farmers over the period of stress, to enable them to remain on their farms to plant such an acreage as may be determined to be wise under all the conditions, with a view to increase the food supply of the Nation and to add to the national security and defense. It is distinctly not intended to be used to stimulate the planting of wheat or any other grain where such planting is not wise from an agricultural point of view and where other activities are safer. Loans were made only in areas that had suffered two successive crop failures from drouth or winterkilling, and to farmers who had land in condition for sowing and who, by reason of crop failures, had exhausted their resources and were without commercial basis of credit. No loan was to be made to any farmer who had unencum- bered real or personal property sufficient to secure a loan of $300. The maximum loan to an individual was $300, and the maximum amount loaned per acre was $3.00. Loans were made on fall wheat and rye, and were secured by mortgages on those crops. Applicants for loans agreed to use seed and methods approved by the Department of Agriculture. These conditions were agreed to by representatives of the Department, the State college of agriculture, and the extension division. They were the minimum which it was felt must be provided if the farmer was to do his part in attempting to produce a crop, rather than the ideal conditions. For example, the requirements for the sowing of fall wheat and rye in Montana were as follows : 48 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. 1. Plowing. — The land on which the grain is sown must have been well plowed since August 1, 1917. The plowing must have been sufficiently well done so that there is no evidence that the native sod is re-establishing itself. 2. Condition of Seed Bed. — The land should be free from weeds and must be sufficiently free to allow the uniform covering of the seed to the proper depth. Land which was sown to a crop in the fall of 1917 or the spring of 1918 must be mellow enough to allow seeding to a depth of at least two inches. 3. The Seed. — The seed sown must be of good quality and practically free from weed seeds. The seed must be of the hard red winter type of such varie- ties as Turkey Red, Kharkov, Crimean, or Alberta Red and must have been treated for smut either with formaldehyde or bluestone solution. The seed of both wheat and rye must germinate at least 80 per cent. 4. Date and Rate of Seeding. — The wheat must be sown not later than Octo- ber 15 and the rye not later than November 1, 1918. Not less than 35 pounds of seed must be sown to the acre. The original tentative allotment of the $5,000,000 fund was $1,750,000 to Montana, $600,000 to North Dakota, $1,250,000 to Kansas, $850,000 to Oklahoma, $350,000 to Texas, and $200,000 kept as a reserve fund. Later, on representations from Washington State officials and farmers, $200,000 of the amount originally allotted to Montana was set aside for loans in that State. A small allotment was also made to New Mexico. The farmer who wished to obtain a loan filled out an application blank at his local bank, on which he stated the acreage he wished to sow to fall wheat or rye, the location and number of acres owned by him, and the number and value of each class of livestock. He also showed the mortgages against his real and personal property and the acreages in crops in 191 7 and 191 8, with the yields produced. This was followed by a sworn statement that he was unable to sow the stated acreage without the loan for which he was making application. In addition, a certificate was required from his local banker to the effect that the farmer's statement of his financial condition was es- sentially correct, and that he did not have sufficient basis for credit to secure a loan of $300. This was followed by a certificate from the local Farm Bureau regarding the applicant's credit condition and his reputation as a farmer and citizen, and also showing whether or not he had the stated acreage in condition for sowing. When the applica- tion was thus complete, it was forwarded to the central office, where is was examined and, if satisfactory, approved by the designated representative of the Department of Agriculture. It was then turned over to an official of the Federal Land Bank, who issued a certificate of approval. This certificate stated that the application for the loan had been approved, and that the money would be advanced by the WARBURTON '. FEDERAL SEED LOANS. 49 Federal Land Bank when a certificate of planting was issued by the county agent and the farmer had signed a note, mortgage, and guar- anty fund agreement. As the only security given by the farmer was a crop mortgage, it was essential that there be no question regarding the legality of this document, and that it be a first lien. For this reason loans were made only to land owners and homesteaders, as renters could not give an unconditional mortgage on their crops. Further, officials of the Treasury Department ruled that a legal mortgage could not be given on the crop until it was actually in existence, and that therefore the mortgage should not be signed nor the funds delivered to the farmer until the crop was sown. This condition was the cause of consider- able criticism and, in occasional instances, worked hardship on the borrower, but local banks very generally accepted the certificates of approval as security for temporary loans to enable the farmer to pur- chase seed and sow the crop. As soon as the seed was sown a planting certificate was signed by the county agent and delivered to the appli- cant or to his banker. The banker then prepared the note, mortgage, and guaranty fund agreement for the borrower's signature, and when this was obtained these papers and the planting certificate were for- warded to the Federal Land Bank. When these all were in proper form remittance was made promptly. The note was an ordinary promissory note bearing 6 percent inter- est. Those in the southwestern district were due October I, 1919, and those in the northern district on November 1, 1919. The mort- gage was a crop mortgage covering a specific acreage of wheat or rye, or both, sufficient to secure the amount of the loan at the rate of $3.00 per acre. The guaranty fund agreement was an agreement by which the borrower obligated himself to pay certain additional sums to the Federal Land Bank if the average acre yield of the acreage mortgaged exceeded 7 bushels. This fund was to secure the Government against loss due to crop failures and consequent inability of borrowers to pay their notes. An average yield of less than 5 bushels per acre was con- sidered a crop failure under the terms of this agreement. Payments into the fund were to be at the rate of 15 cents per acre for each bushel by which the average acre yield exceeded 6 bushels, with a maximum payment of 75 cents per acre. Thus, if the average yield was 7 bushels, the payment was 15 cents per acre; if 8 bushels, 30 cents; if 9 bushels, 45 cents; if 10 bushels, 60 cents; and if 11 bushels or over, 75 cents. Any portion of this fund which remained after the losses occasioned by crop failures and consequent nonpayment were $0 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. adjusted was to be returned to those who had made guaranty fund contributions, each to receive his pro rata share according to the amount of his contribution. Naturally, the county agent and the officers of the Farm Bureau could not be expected to know all the applicants nor to be familiar with conditions on their farms. This was particularly true in some of the larger counties. For instance, Dawson County in eastern Mon- tana measured 170 miles from east to west and 60 miles from north to south. This county was divided early in 1919, but the county agent has continued to look after seed grain loan matters over all his orig- inal territory. To aid county agents and Farm Bureau officials in get- ting first-hand information, the Farm Bureau community committee- men were used. Forms were prepared on which they reported regarding the conditions on the farm of each applicant and the recom- mendations of the Farm Bureau Committee on the borrower's appli- cation were based on this report. Later, the Community Committee reported on the seeding, and on this information the county agent issued planting certificates. Abundant rains in the southwestern district and generally favor- able conditions for sowing caused a heavy demand for seed grain loans in Kansas, Oklahoma, Texas, and New Mexico, and a total of 8,806 applications were approved, totaling $2,025,262. In the north- ern district the moisture was generally insufficient for the best growth of fall-sown crops and the demand was much less than was antici- pated, only 1,837 loans amounting to $371,788 being approved. This was due to the fact that North Dakota farmers generally were able to finance the sowing of rye and that there was no need for financial aid in the better fall wheat sections of Montana. There was practi- cally no demand for loans in Washington, the $300 maximum being regarded by farmers in that State as too small to be of benefit. After it became evident that only a small portion of the amounts allotted to the northern States would be used for loans on fall wheat and rye, arrangements were made to use the remainder of the fund for financing spring wheat seeding. About October 1, 191 8, announce- ment was made that funds allotted to North Dakota, Montana, and Washington and not used for loans on fall wheat and rye would be available for loans on spring wheat at the rate of $5.00 per acre. If a loan had been obtained on 100 acres of fall wheat or rye, no loan on spring wheat was made. If a loan had been made on less than 100 acres of fall wheat or rye, the applicant could obtain a loan on the difference between 100 and his fall-sown acreage. If no fall loan had WARBURTON '. FEDERAL SEED LOANS. 51 been obtained, the limit on spring wheat was $500. The amount of the fall loan was fixed at $3.00 per acre because this amount was more than sufficient to purchase seed, and the borrower who put in fall grain could very generally arrange to leave his farm after the seeding was completed and obtain work at good wages to support his family during the winter. On the other hand, the borrower who put in spring wheat would almost necessarily have to remain on the farm from seeding until harvest, so that the increase from $3.00 to $5.00 per acre was necessary to provide funds for subsistence. No loans on spring grain were made in the southwestern district. The demand for loans in the State of Washington was again very small, while the North Dakota demand was well within the allotment. In Montana, however, the call for funds was even greater than had been anticipated, and it was necessary to transfer $600,000 from unal- lotted funds and funds allotted to other States and unused, in order to grant all deserving requests. The number of approved applications for loans on fall wheat and rye and spring wheat, with the amounts of the loans in each of the several States, are shown in Table 1. Table i. — Number and amount of approved seed grain loans by States. State. Winter wheat and rye. Spring wheat. Number. Amount. Number. Amount. 338 $ 65,994 1.356 $ 484,217 Montana 1,481 301,159 5-295 1,850,285 18 4.635 40 12,895 Kansas 3.531 943-147 Oklahoma 3.852 773.271 1.336 292,651 87 16,193 Total 10,643 $2,397,050 6,691 $2,347,397 The procedure in making loans on spring wheat was the same as for making loans on fall wheat and rye. The experience gained dur- ing the previous fall, however, made the handling of the applications much easier, and the prompt payments made by the Federal Land Banks when the completed papers reached them added greatly to the confidence placed in the project by local bankers. The taking of ap- plications for loans on spring wheat was completed well before the seeding season, and in general the crop was sown under favorable conditions. It was everywhere conceded that never before had farm- ers in the dry-land section made as great an effort to produce crops as they did in the fall of 1918 and the spring of 1919. 52 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. While approximately 800,000 acres of fall wheat and rye and 470,000 acres of spring wheat were mortgaged to secure the loans in the various States, the actual acreage seeded from the proceeds of the loan was considerably greater. No restriction was placed on the use of the funds other than the sowing of an acreage sufficient to secure the loan. In many cases the entire amount of the loan was used in the purchase of seed, so that the acreage sown was considerably greater than was needed to meet the requirements. The placing of approxi- mately $5,000,000 of Federal funds in the drouth-stricken areas added materially to the confidence of bankers and other business men and made it much easier for farmers who had some commercial basis of credit to obtain funds to finance their seeding. As a consequence, the area sown to wheat and rye for the 1919 crop was increased perhaps as much as 2,000,000 acres by the making of Federal seed grain loans. The early spring rainfall in Montana was sufficient generally to enable farmers to put their land in good condition for seeding and to insure prompt germination. In the triangle from Great Falls to Havre and west to the Continental Divide, as well as in portions of Fergus County, the rainfall was insufficient for germination and early growth, while in some portions of eastern Montana and generally in North Dakota heavy rains during the latter part of April and early May delayed seeding. A hot, dry period over Montana and North Dakota began about May 15, with temperatures during the latter half of May running to nearly 100 F. June and July temperatures were also far below normal, while generally the rainfall was below normal. For instance, the average mean temperature in Montana during June exceeded the normal by 4.5 °, while the average precipitation for the State was 0.66 inch, as compared with a normal of 2.80 inches. In North Dakota the mean temperature was 4.4 above normal, and the rainfall was about two-thirds of the normal. As a result, the tillers on rye in western North Dakota and Mon- tana and on fall wheat in Montana were generally killed and only the central stems produced heads. Even these were dwarfed in many sections and matured little or no grain. A considerable part of the rye crop was cut for hay. In some localities showers which fell about July 1 benefited the spring wheat crop, but the average yield of wheat in the sections of Montana where seed grain loans were made and in North Dakota west and south of the Missouri River probably did not exceed 2 bushels to the acre. North of the Missouri River in North Dakota the conditions generally were more favorable and fair yields were produced, altho even here the extreme heat cut the yields to half those which were in prospect shortly before harvest. WARBURTON : FEDERAL SEED LOANS. 53 In the southwestern district conditions were very favorable dur- ing the fall of 1918 and the suceeding winter. The snowfall was heavy, and there was practically no winterkilling. In New Mexico, Texas, and Oklahoma the crop matured under favorable conditions and yields well above normal were obtained. Heavy rains in western Kansas preceding and during heading caused a rank growth of straw and favored the development of leaf rust and other diseases. Lodg- ing and rust, together with a hot, dry period which came during the latter part of June, caused the yields of many fields to fall far below the early estimates, but in general the returns in Kansas were fairly satisfactory. As this is a meeting of agronomists, it is in order to discuss briefly the conditions which led to the widespread crop failures and their probable effect on the future development of the two districts. The southwestern district, which includes the western half of Kansas, the western half of Oklahoma, the Texas Panhandle, and a few counties in northeastern New Mexico, embraces much territory not usually considered safe for the growing of wheat. Because of the high price of wheat in 191 7 and 1918 and also because of a patriotic desire to increase food production, farmers in this district greatly increased their wheat acreage even tho weather conditions at seeding time in 1916 and 1917 were distinctly unfavorable. The available moisture was insufficient for normal germination and growth and, as a consequence, losses from soil blowing and winterkilling were ex- cessive. The drouth continued during the growing seasons of both 1917 and 1918, so that the comparatively small acreage which sur- vived the winter produced small yields. The high price of wheat led farmers who usually devoted considerable acreages to feed crops such as corn and the grain sorghums to increase their wheat acreages ma- terially at the expense of these crops and the usual proportion of fallow. The winterkilling of wheat left land vacant for feed crops, but the rainfall was insufficient for normal yields. The good returns from wheat in 1919 have enabled farmers in the Southwest generally to recoup in large part the losses incurred during the previous two years. The indications now are that the funds re- ceived from this wheat crop are being used for the payment of debts, the purchase of livestock, and for other purposes which will result in lasting benefit to the agriculture of this district. Altho conditions were again favorable for the sowing of wheat over most of the south- western territory in which loans were made, reports indicate a decided reduction in the acreage as compared with that sown a year ago. 54 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. This is due to the exorbitant wages demanded by harvest hands, to difficulties and losses met in marketing the crop because of an insuffi- cient supply of cars for shipment, and to uncertainty regarding the price of the 1920 crop. It is due also in large part to the very general campaign for conservatism in the sowing of wheat and for an increase in the acreage of sorghums and other drouth-resistant crops suitable for feeding to livestock. In western North Dakota and in Montana, farmers had also strained every resource to sow the maximum acreage of food grains during 1917 and 1918. The conditions caused by the general drouth in western North Dakota and in northern Montana during these two years were aggravated by serious losses from the rust epidemic which greatly damaged the wheat crops in western North Dakota and in a few counties in eastern Montana in 1916, and by grasshopper damage, particularly to flax, in western North Dakota in 191 7 and 191 8. A considerable part of the affected area in Montana was public land recently opened for settlement, so that in many cases farmers had not yet completed proof on their homesteads and therefore were unable to obtain funds by mortgaging their land. These people had settled in Montana in 1916 or 1917 and had not produced a profitable crop since their arrival in the State. Others who came earlier had been misled by the abundant crops harvested in 191 5 and 1916 and, on the assumption that yields of 40 bushels or more of wheat could be ex- pected each year, had expended the proceeds in improvements on their land, in the purchase of additional acreages, or in other ways, so that they had little to fall back on when disaster came. For the first time since weather records have been kept in western North Dakota and in Montana, three dry years have occurred in suc- cession. Not only have farmers suffered from the cumulative effects of the drouths of 1917 and 1918, but the drouth in 1919 was more severe and widespread than that of either previous year. Tho the average rainfall in this region is less than in western Kansas and west- ern Oklahoma, the normal summer temperature and the evaporation are usually much lower, so that the precipitation is decidedly more effi- cient. In the judgment of men who know conditions in both sections, the chances of producing a profitable wheat crop in Montana and western North Dakota are better than in western Kansas and western Oklahoma, but this year the farmers in Kansas and Oklahoma gam- bled with the weather and won, while those in Montana and south- western North Dakota gambled and met a most disastrous loss. As a result, a considerable number of the farmers have been com- BUCKMAN : TEACHING ELEMENTARY SOILS. 55 pelled to leave their farms temporarily and secure employment else- where to support their families. Much of the livestock has been sold, and the purchase of feed and forage at high prices has been necessary to maintain that which was retained. The courage of the farmers, however, is remarkable, and only a very small percentage is leaving the farms permanently. The successive drouths have caused great interest in irrigation, particularly in Montana, and probably the irri- gated acreage will be greatly increased within the next few years. An increased irrigated acreage will not only stabilize Montana agri- culture, but will be of great benefit to the dry farmers, as it will insure an abundant supply of forage for use in dry years and a market for surplus stock. In my opinion, dry farming has a future in western North Dakota and Montana, but it must be based largely on farm units of at least 640 acres and an increased production of livestock. For the most part this livestock will be pastured on native range, but will be main- tained thru the winter and thru periods of temporary drouth by the use of forage crops such as corn, sunflowers, and sweet clover. A considerable acreage of wheat and other small grains will continue to be grown, but the main dependence of the farm will not be on these crops. A discussion of Federal seed grain loans would not be complete without an acknowledgment of the excellent cooperation rendered by the various State extension officials, county agents or other county officials who acted in their stead, officers and members of farm bureaus, and local bankers. The county agents in particular are to be commended for their untiring efforts in connection with these loans. Without their aid, proper supervision in making the loans would have been practically impossible. THE TEACHING OF ELEMENTARY SOILS. 1 H. O. BUCKMAN. The teaching of soils in the systematic and fundamental manner in which recognized university and college subjects have long been handled is a comparatively recent development and as a consequence the possibilities are as yet only partially realized. As with all new subjects, soil science is going thru a definite evolution and will con- 1 Contribution from the Department of Soil Technology, Cornell University, Ithaca, N. Y. Received for publication December 4, 1917. 56 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. tinue to so evolute until it is placed on as sound a theoretical and pedagogic basis as its subject matter will permit. In general, new- sub jects lack both data and arrangement. Soil science is no excep- tion to the rule. At the beginning, little scientific knowledge was available and even that little had no logical sequence. Taught in many places as a part of agricultural chemistry, no distinct laboratory de- velopment resulted, especially regarding the physical phenomena. The need for teaching data has been acute and while this need stimulated research, it forced the teacher for the time being into a very practical and applied consideration of his subject. Lectures, demonstrations, and field explanations constituted most of the in- struction. The agricultural colleges were growing rapidly and expan- sion was expected at every point. The teachers of soil technology felt the pressure and as most scientific subjects are taught with lab- oratory practice, a move in that direction seemed desirable. The chemical phases were, of course, developed with chemical equipment as a basis. The physical side of the work was, however, much handi- capped by lack of apparatus and methods. Due to the difficulty of controlling physical reactions in the soil, the exercises were always unsatisfactory and often crude. To a student with a good knowl- edge of elementary physics such laboratory practice was not only a waste of time but almost an affront to his general education. The chemical phase of the laboratory took a different turn, and in- volved analytical chemistry was often taught in an attempt to give the student a chemical viewpoint of soils. Time was wasted in chem- ical methods and the attention thus distracted from considerations invaluable in understanding the " why " of practical farm operations. Chemical principles that may apply to soils are a part of the science of chemistry and should be taught in a department of chemistry and not in a department of agronomy or soils. Free use of such prin- ciples, however, is the right of every soil teacher. The time has come when many feel the necessity of revising the teaching methods which have developed during the infancy of the subject. Plenty of sound data is at hand for such courses. It is not a question of knowledge but a question of arrangement and co- ordination. The pedagogy of the subject so long neglected needs attention. How to teach is the pertinent question, not what to pre- sent. Recent scientific and theoretical advancements have a proper place in the more technical courses offered by a department. The science of the soil is now on such a basis that one general fun- damental course seems preferable to the two or even three that are BUCKMAN : TEACHING ELEMENTARY SOILS. 57 in many places offered in as many calendar terms to cover the subject. Good pedagogy demands such a change. Geology has long been taught in such a manner, as have other pure sciences. Is soil science so different that ordinary procedure does not apply? Has it such a large body of facts that three lectures a week for a term will not permit a clear presentation of the fundaments ? The institution of recitation periods should be the next step. Here the principles explained and emphasized in the lectures can be ex- panded and discussed, preferably with a textbook as a basis. The old style laboratory " experiments " should be discarded and exercises substituted which emphasize fundamental points. In short, the whole course may be made a " follow-up," the recitations on the lec- tures and the laboratory on both. Few ideas not already explained in lecture or recitation should be introduced into the laboratory. The student should there be given a chance to handle, study, and experi- ment with the material previously discussed. The study of soil- forming rocks and minerals, the naming of soils, the estimation of organic matter, the testing for acidity, and the identification of ferti- lizers are only a few of the possible follow-up exercises. Such a cor- relation of lectures, recitations, and laboratory could not fail to raise the grade of the instruction, especially if the students are properly grounded in chemistry, general geology, and physics before register- ing for the course. At present, the institutions that are seriously engaged in teaching soils are not closely enough in touch so that any of them realizes the problems which are confronting the others. Few opinions and views have been exchanged. The experiences of one institution, whether successful or otherwise, have gone for naught as far as the others are concerned. As a consequence, a conference for a thoro discussion of the points at issue seems not only desirable but almost a necessity. Such a conference would aid the science as well as the pedagogy of the subject. 58 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. INTRODUCTORY COURSES IN SOILS. 1 R. S. Smith. In a recent number of this journal, Karraker 2 raised a number of interesting questions regarding the laboratory work usually offered in the first or general course in soils by our State agricultural colleges. The following specific criticisms are implied : The work is not suffi- ciently vocationalized ; it lacks the qualities for imparting mental dis- cipline most effectively ; it lacks accurate control ; and the results obtained in certain exercises are misleading. It is further stated that many of the fundamentally important processes cannot be reproduced in the laboratory and the question is asked whether the student should not be given the opportunity of getting the subject matter of the course without the laboratory work. The present writer feels that efforts directed towards the improve- ment of any course, particularly an introductory one, should be based on a careful examination of the true purpose of the course as a whole. A less comprehensive effort is likely to fall short of attaining the maximum possibilities because of the lack of a clearly defined aim. The purpose of an introductory course may be tentatively stated to be not merely to impart information, but also to enlarge the stu- dent's vision and stimulate and illuminate his thinking. The full accomplishment of this purpose even by a well-equipped and enthu- siastic instructor who has an abiding faith in the value of the work he is giving is probably impossible, but it may be attained with more or less success by interspersing thruout the strictly instructional work glimpses into the history of the subject, appreciative mention of some of the men who have contributed to it in the past and of those who are now contributing, and by calling attention in a conservative way to some of the unsolved problems which still challenge the best abili- ties. The effort should be to inspire the student to enter new worlds of thought with enthusiasm and to give him an understanding of the fundamental principles which underlie successful practice and not simply teach him to handle soils so they will grow good crops. 1 Contribution from the Division of Soil Physics of the Department of Agronomy, College of Agriculture, University of Illinois, Urbana, 111. Received for publication November 21, 1919. 2 Karraker, P. E. What is the value of the usual laboratory work given in general soils courses? In Jour. Amer. Soc. Agron., 11, no. 6, p. 253-256. 1919. SMITH : INTRODUCTORY SOILS COURSES. 59 If the admittedly weak laboratory exercises in introductory soils courses are to be strengthened by revision, elimination, or addition in the most rational manner, their contribution to the attainment of the purpose of the work as a whole must be kept in mind and they must be made to conform to this purpose as well as to the best available knowledge. An examination of the true function of each part of a course must be made and efforts at the improvement of any part must be guided by a clear idea of its function. As ordinarily given, the introductory soils course consists of two or three lecture periods, one recitation period, and one or more labora- tory periods a week. The lecture work is the center or nucleus upon which the other parts hinge and affords better opportunity for stim- ulating interest than is afforded when it is replaced by textbook as- signments and an increased number of recitations, tho it may not be as effective a method of imparting information as the latter plan. The recitation period is designed to aid the student in making the in- formation presented in lecture or textbook a part of his experience. It is difficult to carry the inspirational features often present in the lecture work thru the recitation period, for its nature seems to demand a rigid insistence that the facts previously presented be assimilated by the student. The higher the scholastic quality of the class, the easier it is to measure up to the general purpose of the course, as above outlined, in the recitations. It seems to the writer that it is unfortunate to tenti or think of the recitation hour as a " quiz " period. It should rather be emphasized as a discussion period, tho this does not imply the aimless and time-consuming discussion so dear to the heart of the lazy student. The laboratory work is designed as a further aid to the student in visualizing and making real the fundamental principles brought out in the class room. It is exceedingly difficult to correlate the labora- tory work properly with the lecture or textbook work, particularly in soil physics. Failure to accomplish this correlation results in the stu- dents performing the set exercises simply as a job to be done without an appreciation of their relationships. A confused state of mind re- sults on the part of the good students and an attitude of indifference on the part of the poor ones. It seems fundamental to the writer that material which has not been treated in lecture or in textbook assign- ments should not be introduced in the laboratory, because of the ap- parently unavoidable confusion resulting. Whether this ideal can be fully attained in the soil physics course is a serious question, particu- larly when several periods a week are devoted to laboratory work. Altho the difficulty is, great, its solution offers ample reward. 60 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. The subject matter of the laboratory exercises in soils courses has been questioned by Mr. Karraker. So far as the exercises lack ac- curate control or lead to false conclusions they are to be severely criticised. That they are not sufficiently vocationalized and fail to give the student " a knowledge and something of the art of doing operations which he will be using in post-graduation activity " does not appear to be an equally valid criticism. The educational value of any part of an introductory course cannot be judged by such a cri- terion, but must be judged on the basis of its contribution to the pur- pose of the course as a whole. If any given part cannot harmonize with and make a definite and real contribution to this general purpose, then the question may be legitimately asked whether this part should not be eliminated at least for certain groups of students. The problems presented by the first soils course are difficult of solution. They are becoming constantly and rapidly more difficult, due to changes in the nature of the courses given in secondary schools. The need for improvement is general and must be met. The purpose of this paper has been to state in broad terms a ten- tative outline of the general purpose to be attained by the introductory soils course and not to present a criticism of the many admittedly weak points. The need seems to be for something more than the im- provement or elimination of certain laboratory exercises. Individual effort is always limited by the peculiar experiences and idiosyncrasies of the individual. Basic principles must be worked out and, if pos- sible, agreed upon. To accomplish this the best collective effort of the men who are responsible for the teaching in the introductory soils courses is necessary. Efficient machinery must be set in motion for the definition of principles and correlation of the parts. The writer ventures to suggest that an effective and feasible method of undertaking the improvement of our introductory soils courses which appears to be so generally needed would be for the American Society of Agronomy to select a group of men and instruct them to undertake a thoro study of the entire problem and make such recommendations as they saw fit. Such a group should define as clearly as possible the fundamental object to be attained by the courses and should also make a minute study of the means now em- ployed or possible of employment in attaining this object. Great good would certainly result from the efforts of such a group in giving rational direction to the efforts of the instructor who may now feel almost lost among the growing difficulties with which he labors. CONNER! EFFECT OF ZINC IN SOIL TESTS. 6l THE EFFECT OF ZINC IN SOIL TESTS WITH ZINC AND GALVANIZED IRON POTS. 1 S. D. Conner. During the progress of some pot tests with acid soils, some unex- pected results were obtained with the crops grown. After the first year, the unlimed soils became less productive. For instance, soils which grew fair clover the first year later would not grow buckwheat, a crop known to be much more tolerant toward acid soils. The pots used were of zinc and galvanized iron, and were well paraffined to prevent the action of the soil acids upon the zinc. The results obtained, however, raised the question as to whether or not zinc had been made soluble and had penetrated the paraffme coating. With this point in view, tests were made and zinc was found in quite large quantities in the unlimed soils but not in the limed soils. HISTORICAL. In 1904, Veitch (8) 2 reported the presence of zinc in the sodium chloride solution of two acid California soils. Upon inquiry as to the character of these soils and the source of the zinc, Veitch later made the statement that the zinc came from pots and was not naturally found in the soils. Haselhoff and Gossel (5) in 1904 reported that zinc sulfate was highly injurious to wheat and that the injury was not reduced by liming. Zinc oxide was not found to be injurious. This apparent contradiction might possibly be explained on the basis of a different degree of distribution of the zinc in the soil. The zinc sulfate, being soluble, might be much more widely distributed before fixation even in limed soil than was the insoluble zinc oxide. Meyer in 1905 (6) found that crop yields were much greater in earthenware than in zinc pots, and that the injurious effects were largely overcome where lime carbonates were. used. Tacke (7), in 1 Contribution from the Department of Soils and Crops, Purdue University Agricultural Experiment Station, La Fayette, Ind. Read at the twelfth annual meeting of the American Society of Agronomy, Chicago, 111., November 11, IQIQ. 2 Reference is made by figure in parentheses to " Literature cited," p. 64. 62 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. commenting on Meyer's article, says that the injury was heightened by acids dissolving zinc from the pots and that he had the same ex- perience when experimenting with acid moor soils. Ehrenberg (i) in 1908 reported work where he buried zinc plates in the soil. The zinc was found to be injurious, and ammonium salts caused increased injury. This would be in accord with the known action of ammonium sulfate in increasing soil acidity. Ehrenberg (2) suggested that different results with organisms in field soils and in zinc pots might be due to the interfering action of zinc salts. In 1910, he reported (3) beneficial results of zinc salts where nitrate of soda was used and injurious effects where ammonium salts were used. Voelcker (9) in 1913 reported pot culture experiments where he found that zinc up to 0.01 percent in the soil was stimulating to wheat and above that was toxic. Ghedroiz (4), in reporting experiments with acid soils in zinc pots, stated that clover made feeble growth in the third year. The bad effects were increased in the absence of lime. Clover grown in ebonite vessels showed no injury. EXPERIMENTAL DATA. Samples of five different types of acid soils which had been kept in paraffined galvanized pots for two and a half years were taken for the purpose of determining the percentage of zinc in both the un- treated soils and from pots which had been heavily limed. The results are shown in Table I. Table i. — Percentage of sine in acid soils from paraffined pots, with and with- out calcium carbonate. Without lime. With lime. a Soil. Volatile matter. Acidity, Acidity, KNO3 Zinc. c KNO3 Zinc.c method. 1 method. 6 Percent. Pounds. Percent. Pounds. Percent. Yellow clay 3-57 2,800 0.021 None 0.001 White silt 3-92 1,800 .013 do. .002 7-45 840 .014 do. .002 10.13 2,520 .028 do. .001 Brown peat 85.20 9,600 .088 600 .010 a The CaC0 3 added per 2,000,000 pounds of soil was, for the white silt and brown loam, 6 tons each ; for the yellow clay and black sand, 12 tons ; and for the brown peat, 40 tons. 6 Pounds of CaC0 3 needed per 2,000,000 pounds of soil. c Zinc soluble in cold normal KN0 3 solution. CONNER: EFFECT OF ZINC IN SOIL TESTS. 63 In each unlimed soil, there was more than 0.01 percent zinc soluble in cold normal KNO s solution, the amount Voelcker (9) had found to be injurious to wheat in a loam soil which contained over 1 percent' of lime. In no instance did a neutral sample of soil contain zinc in more than mere traces. The brown peat, which is of an excessively acid type, contained the most zinc and this soil, to which 40 tons of lime per 2,000,000 pounds of soil had beed added, was still acid and contained 0.01 percent soluble zinc, an amount approaching the danger line of toxicity. Table 2 shows the relative yields of wheat and clover grown the first season and the yields of buckwheat and oats grown the last season on the two least acid of the five soils. The yields from the other soils are not given because they were very small, even with the first crops, on account of high acidity on the unlimed soils. On these soils fair yields of wheat and clover were obtained, in pots where smaller amounts of lime were applied, and later the buckwheat and oats failed almost entirely, undoubtedly because of soluble zinc. Table 2. — Average yields in grams of wheat and clover in igiy and buckwheat and oats in 1919 grown in paraffined galvanized pots in acid and limed soils. ■ Soil. First year. Last year. Wheat. Clover. Buckwheat. Oats. White silt 20.0 15.0 0.2 O.I White silt limed 41.O 40.0 31-2 1-7 20.0 2I.O 3-5 i-7 Brown loam limed 27.O 32.0 35-0 2.0 Whether a good protective coating can be found is a question re- maining unsolved. 3 Until a satisfactory coating is found there can be no question that it is unsafe to use zinc-coated pots in soil tests with acid soils. The pots used in these experiments were 9.25 inches in diameter and were watered by a subirrigating device on the order of the Wagner pot, so any injurious substance would have a chance to penetrate the soil quickly. W'here large pots or rims are used and the water is applied at the surface it would undoubtedly take longer for the zinc to permeate the whole soil mass ; hence, the injurious effects would be delayed. . 3 Dr. H. J. Wheeler has given the information that at the Moor Culture Ex- periment Station at Bremen, Germany, zinc pot's could not be used on account of the soil acids until they had received a protective coating. Wood tar applied with heat was the most successful material used. 6 4 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. In this connection it might be pertinent to point out that this action of acid soils on zinc is one more bit of evidence that soil acidity is chemical and not physical and that there are real acids in the soil and not mere selective absorption phenomena. SUMMARY. 1. In pot tes^s with acid soils in paraffined galvanized iron pots, crops began to fail the second season in an unexpected way. 2. On testing the soils where the crops had failed, water-soluble zinc salts were found. 3. Tests of the soils showed that the zinc was present in injurious amounts in the untreated or insufficiently limed soils, but not in the soils where sufficient lime had been used. 4. The paraffine coating had been granulated and the soil acids and zinc salts had passed thru it. 5. No good protective coating has yet been found. 6. This action of acid soils on zinc is evidence that soils contain true acids. LITERATURE CITED. 1. Ehrenberg, P. The action of lime in pot experiments. In Chem. Ztg., v. 32, no. 78, p. 937. 1908. 2. . A review of the bacteriology of ammonium compounds. In Fiihling's Landw. Ztg., v. 57, no. 13, p. 449. 1908. 3. . The action of zinc in pot experiments. In Landw. Vers. Stat.; 72 : 15-142. 1910. 4. Ghedroiz, K. The influence of zinc vessels in culture experiments. In Selsk. Khos. i Lyesov., 74: 625. 1914. 5. Haselhoff, E., and Gossel, F. The effect of sulphurous acid, zinc oxid, and zinc sulphate on plants. In Ztschr. Pflanzenkrank., v. 14, no. 4, 193 p., 2 pi. 1904. 6. Meyer, D. The injurious effects of gypsum in vegetation experiments in zinc pots. In Fiihling's Landw. Ztg., v. 54, no. 8, p. 261-267. 1905. 7. Tacke, B. On the injurious action of gypsum in vegetation experiments in zinc pots. In Fiihling's Landw. Ztg., v. 54, no. 10, p. 331, 332. 1905. 8. Veitch, F. P. Comparison of methods for the estimation of soil acidity. In Jour. Amer. Chem. Soc, 26: 637. 1904. 9 Voelcker, J. A. Pot culture experiments. In Jour. Roy. Agr. Soc. England, 74:411. 1913. FIPPIN : THE TRUFAST TEST. 6 THE TRUFAST TEST FOR SOUR SOIL. 1 Elmer O. Fippin. The practical experience of farmers and the investigations of State and Federal agencies show that large areas of soil in humid regions need application of liming materials if their full productive capacity for many important crops is to be attained. The chemical, physical, and physiological significance of this need for lime in the soil has not been made clear by the investigations con- ducted thus far. However, there seems to be a fairly good agreement among investigators that this need for lime is usually associated with, if not directly due to, a considerable degree of acidity in the soil. Acidity in this sense is perhaps best defined as the capacity of a soil to absorb calcium, sometimes called the lime absorption coefficient. Whether the capacity of the soil to absorb other bases than calcium bears a direct relation to its need for lime does not seem to have been determined. It is well known that soils differ widely in their need for lime for the successful growth of some crops. The prevailing practice has been to associate this different need for lime with the presence of different amounts of active or free acid in the soil. Various tests are in use to measure the amount of this so-called acidity which is as- sumed to be correlated with the amount of lime needed by the soil for the production of crops sensitive to lack of lime in the soil. From the point of view of farm practice such methods of measuring the need for lime in the soil should be rapid, simple, and easily carried out either in the field or in close association with field conditions. Without presuming to review the essential features or the advan- tages and disadvantages of existing methods for this purpose, we pre- sent a new method that has been devised to meet, as far as now seems possible, the dominant requirements of the problem of so-called acid or sour soils. This test is called the " Truf ast " test, the manipula- tion and most of the essential features of the test having been devised by Mr. A. D. Whipple, who has been associated with the National Lime Association as its chemist and engineer. The underlying chem- 1 Paper 3-20-2 from the National Lime Association, Washington, D. C. Pre- sented before the twelfth annual meeting of the American Society of Agronomy, at Chicago, 111., November 10, 1919. 66 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. ical principles employed are much the same as in other tests. The most distinctive features are the chemicals employed and especially the outfit and manipulation by means of which the test is made. CHEMICAL PRINCIPLES EMPLOYED. To bring the acidity of the soil into liquid suspension where it can be measured, the soil is suspended in a strictly neutral solution of calcium nitrate of sp. gr. 1.3 at 76 F. After suspension in 60 c.c. of this solution for five to ten minutes with thoro shaking, the mix- ture of soil and solution is filtered through a neutral filter paper and the first 15 or 20 drops of extract are rejected to correct for absorp- tion of the paper. An aliquot portion of the soil extract, 15 c.c, is used for titration. The alkali used in titration is a solution of sodium carbonate hav- ing a strength of iV/49.5. The indicator used is methyl red, which does not react freely to carbonates. SPECIAL FEATURES OF THE TRUFAST OUTFIT. The amount of soil employed in the test is 6.166 c.c, which is one hundred-millionth of the volume of an acre 6 inches deep. This is approximately the volume of soil normally treated by an application of lime. The sample of soil is measured by liquid displacement. A specially made cylinder is employed having a mark on the glass at 60 c.c, to which line the cylinder is filled with calcium nitrate solution. Soil is then added until the level of the liquid is raised to the level of a sec- ond line marking a volume of 66.166 c.c, or 6.166 c.c above the first. All liquids are measured from the top of the meniscus to avoid error in turbid solutions. Vigorous agitation by shaking about 500 times is employed to break down granulation and permit free diffusion from the surface of the soil particles. The second or titrating cylinder has a scale cut in the glass, the lower or zero line of which marks a volume of 15 c.c. from the bot- tom of the cylinder. This is the aliquot portion of the extract used for titration. The indicator is placed in the titrating solution in the proportion of about one drop to the cubic centimeter. The soil ex- tract is titrated in the cylinder direct from the bottle, a few drops at a time, until any red color has been discharged and a neutral tint at- tained as shown by comparison of the color of the titrated extract with the color of the titrating solution when both are viewed against the light. At this point it is important to stop the titration as soon as FIPPIN : THE TRUFAST TEST. 67 the last trace of red color disappears from the titrated column of soil extract as viewed laterally against the light and not attempt to dis- charge the residual pink color that may still be seen from the top of the solution column. The complete discharge of the end point color represents a strongly alkaline condition of the solution and not the neutral condition that is desired. The difference in depth of solution in the cylinder and the bottle must be taken into account in making this color comparison. The number of cubic centimeters of alkali required is read direct from the scale, the strength of the alkali in comparison with the vol- ume of the soil used being such that each scale division represents a lime absorption equivalent to 500 pounds of calcium oxide per acre 6 inches of soil. Every fourth division represents 1 ton and these are numbered from the bottom. A test is ordinarily made in from 10 to 15 minutes. Samples will vary in the rate at which they filter. In running a series of samples the filtering and titration can usually be done as rapidly as the sus- pension of soil is prepared. In using calcium as the absorbed base the test is comparable with the liming material the farmer will use. Nitric acid is a normal constituent of soils and is only very slightly absorbed. In measuring the soil by liquid displacement, errors due to crumb structures are largely avoided. No heat is used. The neces- sity for all mathematical calculations is eliminated by the character of the outfit and the strength of the solution. The test reads directly in practical terms. The reading of calcium oxide may be readily con- verted into hydrated lime by multiplying by the factor 1.35 and to carbonate forms of lime by using the figure 1.8 for pure materials or more for material of coarse texture or that is impure. Clear solutions are secured. The results of the test on soils rich in organic matter are consistent with the results in normal mineral soils. No attempt has been made at an extended comparison of Trufast results with the results from other methods. That is left to official laboratories. The results of a few determinations are given in Table 1 to illustrate the character of the readings. Table i. — Results of determinations of soil acidity by the Trufast test. Trufast reading, pounds of calcium oxide per acre 6 inches. Type of growth. I. 2,300 Good clover 2. . . . 2,700 do. 3- Cecil clay soil, Md . . . 3,ooo Medium clover 4- Hagerstown silt soil, Md ... 1,500 White clover inoculated 5- Volusia silt loam, N. Y . . . 4-500 Clover scarcely grows 6. Ontario loam, N. Y . . . 2,500 Good clover 68 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. These results give some idea of the character of the reading se- cured. In my judgment it is not reasonable to expect an exact agree- ment between the results from different methods that may be em- ployed in view of the complicated chemical character of the soil. Further, it is important to recognize that the readings from any method have not been shown to correspond directly with the amount of lime required by that soil for the best growth of any particular crop. The correlation of crop growth with the coefficient of lime absorption or acidity in a soil is a distinctly separate operation from determining the position of a soil on the acidity or lime-absorption scale. Methods of measuring acidity determine only the position of the soil on such a scale. Each crop has a certain range of tolerance of this so-called acidity and also of free lime in the soil and this range appears to be different for different plants. Having determined the coefficient of calcium absorption of the soil the next step is to determine from the acidity tolerance of each crop the amount of lime necessary to bring the soil up to a condition for good growth of that crop, whether it means an alkaline, a slightly acid, or a moderately acid condition. The determination of the range of tolerance of each kind and variety of plant as regards acidity or alkaline or lime condition of the soil, as determined by any standard method, is one of the important pieces of soil investigation yet to be carried out. This requires a method that combines with reasonable accuracy of quantitative meas- urement, such rapid, simple manipulation as will permit the making of a very large number of determinations under all kinds of soil con- ditions, soil treatments, and crop growth. From such a mass of data we may hope to work out important correlations or possibly the lack of correlations with crop growth. The Trufast test is presented as a contribution to the means for working out these correlations if thev exist. WESTOVER & GARVER ! EXPERIMENTAL SILOS. 6 9 A CHEAP AND CONVENIENT EXPERIMENTAL SILO. 1 H. L. Westover and Samuel Garver. The utilization of certain forage crops for silage has been practiced for many years, corn being one of the first crops to be use in this way. In the early days of silage making, little was known regarding the ripening changes that took place in the silo and the conditions necessary for preserving the material properly were not well under- stood. As a result, the material that was put up frequently spoiled. At the same time, there was considerable disagreement among feed- ers as to the real value of silage from a feeding standpoint. As the conditions necessary for the proper preservation of silage became better understood and as a fuller appreciation of its true value as a feed developed, an increased demand for silage naturally followed. This eventually led to attempts to utilize other crops in this manner, particularly in regions where corn does not succeed and also in those sections where considerable spoilage results from unfavorable weather conditions at hay-making time. Undoubtedly there are many plants, in addition to those at present being used, that could profitably be utilized for silage. In order to determine just what plants can be satisfactorily ensiled and under what conditions the best results may be obtained, some type of experi- mental silo is needed that will make it feasible to work with a large number of containers at a minimum cost. In the past, much of the work in trying new crops for silage has been with the farm silo, but this method is open to certain objections, the most obvious of which are the financial losses entailed when the contents spoil, as frequently happens, and the very limited number of experiments that it is pos- sible to conduct. Several State experiment stations have tried various methods for ensiling crops in small quantities and have reached the conclusion that silage in small containers is equal to that in large silos as judged by appearance and chemical analysis, where the conditions of ensiling are properly controlled. The silages made at these stations for ex- perimental purposes fall into two general groups, chopped materials 1 Contributions from the Office of Forage Crop Investigations, Bureau of Plant Industry, U. S. Department of Agriculture, Washington, D. C. Re- ceived for publication January 10, 1920. JO JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. in small silos in quantities sufficient to test their palatability and feeding value in addition to studying the ripening changes, and chopped materials in tightly sealed glass jars and milk bottles for chemical and bacteriological studies, etc. One of the early reports on the use of the larger experimental silos was made by the Wisconsin Agricultural Experiment Station for the \ear ending June 30, 1888. A bay in a barn was divided into six compartments, each 7 by 8 by 14 feet in size, but the materials placed in these silos did not come out in very good condition, and it was concluded that silage would not keep well in small silos. About twelve years later the Oregon station constructed several silos 9 and 10 feet in diameter and 22 feet deep. Other States that have more recently reported on the results obtained with small silos are Kansas, Missouri, and Idaho. The Idaho station used fir stave silos 3 feet in diameter and 6 feet deep, holding approximately 1,500 pounds. The Missouri station used cypress silos of the same capacity as those used by the Idaho station. These silos were fitted with covers, which at the Idaho station were weighted to insure proper settling of the contents, thus more nearly approximating conditions in a large silo. The silos used by the Kansas station were 7 feet in diameter and 16 feet high, with a capacity of about 10 tons. They were covered with a roof and as the contents did not settle satisfactorily the first year, the second season it was weighted with bags of sand to make condi- tions similar to those in the large silos. Among the experiment stations that have reported results with smaller containers are Kansas, Iowa, Connecticut, and Wisconsin. In most cases these containers were 1 -quart or 2-quart glass jars or milk bottles. The Wisconsin station, however, used containers ranging in size from one pint to several gallons, the larger ones sometimes being of galvanized sheet iron. The material was put up under laboratory conditions and was used in studying the ripening processes. It is quite apparent that the larger experimental silos described above could be constructed in large numbers only at an expenditure of considerable time and money. The glass jars, on the other hand, in sizes generally used, are inexpensive, but, while they furnish suffi- cient material for laboratory studies, are of little value in determining the palatability and feeding value of any particular silage. What is really needed is a container that will permit putting up an almost un- limited number of silages at a minimum expense. It is believed that a method employed at the field station of the United States Depart- ment of Agriculture at Redfield, S. Dak., during 1919 meets this re- WESTOVER & garver: experimental silos. 71 quirement. After the experiment had heen completed it was found that the same type of container had heen used hy the Illinois station at least as early as 1889, hut in this case the main ohject of the experi- ment was to study the temperature changes in the silage and the size and character of the container was incidental. About the same time, the Vermont station reported the use of a circular wooden tank '3 feet high and 2 feet in diameter. Both these experiments were conducted so many years ago that they practically have heen lost sight of. The experiment was carried out at Redfield essentially as follows : In August, 1 91 9, a hand-feed cutter was purchased at a cost of about $20. This cutter was adjustable to cut materials in lengths of a half to one and a half inches and for this experiment was set to cut in seven-eighths inch lengths. This chopper met all the demands and not more than half an hour was required to chop enough material to fill each container. The containers w r ere motor oil barrels such as can be purchased at any garage, altho any strongly constructed barrel that is airtight and watertight would answer the purpose just as well. The cost of these barrels will vary somewhat at different times and places, but those purchased at Redfield cost $1.50 each. The heads are easily removed by loosening the two top hoops. The wood seems to have been treated in some way or else it is of such a character that the oils do not pene- trate to any appreciable extent, as the barrels are easily cleansed with water. When the silage was opened it had no evidence of taste or odor of oil. The chopped material was thoroly trampled as it was placed in the barrels and was heaped up somewhat so that it was necessary for a man to stand on the cover in forcing it on. When the cover was in place the hoops that had been loosened in removing it were driven back in place. The head was then covered with a thick coat of paint to exclude the air. In forcing the head in place some sort of a press such as is used in heading apple barrels would doubtless be more satisfactory than the method described above, as it would be possible to get a greater quantity of material in each container. With the method employed at Redfield, the contents of each barrel weighed from 150 to 200 pounds, the weight varying with the material used and the amount of trampling it received. The materials ensiled in this way were alfalfa, sweet clover, corn, sorghum, sudan grass, Russian thistle, wild sunflowers, soybeans, one third corn and two thirds alfalfa, half sorghum and half alfalfa, and 72 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY." half corn and half alfalfa. The alfalfa, corn, sorghum, one third corn and two thirds alfalfa, half sorghum and half alfalfa, and half corn and half alfalfa were put up on August 23 ; the other crops were ensiled on September 3. The barrels were opened on November 12 and 18 and all had a good appearance with the exception of the Rus- sian thistle, which was dark and watery. Altho these feeds were tried on cattle unaccustomed to silage, all were eaten readily except the Russian thistle, which was refused absolutely, and the wild sun- flower, which was eaten very sparingly by one cow. It was apparent that the wild sunflower had not fermented properly. This was not due to the method employed, but probably to the pres- ence of certain resinous substances, as the material retained the strong characteristic odor it had when put in the barrel. All the other silages except the Russian thistle had an excellent aroma and when small quantities were shipped into Washington and placed in glass jars in a warm room several of them remained in good condition for six weeks. The results from this preliminary experiment were so satisfactory and the cost so insignificant that it is believed the method may be used advantageously in testing the keeping qualities of many plants, regardless of their apparent value for silage. There is almost no limit to the number of crops and combinations that may be tried under this method. AGRONOMIC AFFAIRS. MEMBERSHIP CHANGES. The membership of the Society as reported in the January issue was 561. This number included 61 whose membership actually lapsed on December 31 because of non-payment of dues for 1919. After deducting these lapsed members and adding 4 who have joined the Society since the last report, a net loss of 57 is shown, with a present membership of 504. It is hoped that a vigorous campaign for new members will be productive of results, and that many new members will be added during the next two or three months. With present high costs of paper and printing, a satisfactory Journal cannot be published unless the membership is very materially increased. The earnest cooperation of all members is solicited. The names and addresses of new members, the names of those whose membership has lapsed, and such changes of address as have come to the notice of the officers, follow. AGRONOMIC AFFAIRS. 73 New Members. Bryan, Walter E., University of Arizona, Tucson, Ariz. McGinnis, F. W., 2089 Carter Ave., St. Paul, Minn. Odland, T. E., University Farm, St. Paul, Minn. Patrick, A. L., Agronomy Dept., State College, Pa. Agee, John H., Alvord, Emory D., Beavers, J. C, Bennett, Chas. D., Berry, Roger E., Bledsoe, R. Page, Bower, H. J., Brandon, J. F., Bryant, Ray, Bugby, M. O., Burtis, Earl, Bushey, A. L., Cocke, R. P., Cooper, M. L., Cowles, H. C, Daane, Adrian, Emerson, Paul, Firkins, Bruce J., Freeman, Geo. F., Furry, R. L., Gericke, W. F., Lapsed Members. Gilbert, M. B., Hallsted, A. L., Hardenburgh, E. V., Hatcher, Otto, Hildebrand, E. B., Horton, Horace E., Hotchkiss, W. S., HUNGERFORD, DeF., HUNNICUTT, B. H., Iberico, Juan R., Igo, Jerome, Johnson, D. R., Jones, J. W., Kan, F. F., Kelly, E. O. G., Kirk, N. M., Knight, Chas. S., Krall, John A., Kuska, J. B., Lynde, C. J., McHenry, Norris, Mc Miller, P. R., Maier, Fred, Maris, Edwin L., Milton, Roy H., Morgan, J. O., Olson, P. J., Peterson, W. A., Ruzek, C. V., SCHMITZ, NlCKOLAS, SlEGLINGER, J. B., Stokes, W. E., Turner, A. F., Veach, C. L., Voorhees, John H., Waller, Allen G., Walter, E. J., West, J. T., Winters, R. Y., Wyatt, F. A. Changes of Address. Fergus, E. N., Dept. Agron., Univ. of Kentucky, Lexington, Ky. NOTES AND NEWS. D. A. Brodie, who has been connected with the Federal office of Farm Management since 1903, on January 1 became agriculturist for the Western Sulphur Co., with headquarters at Cheyenne, Wyo. F. M. Clement has been appointed dean of the faculty of agri- culture of the University of British Columbia, vice L. S. Klinck, now president of the University. A. D. Ellison, scientific assistant in charge of cereal experiments on the Belle Fourche Experiment Farm, Newell, S. Dak., resigned No- vember 30 to become county agricultural agent of Butte Co., S. Dak. F. V. Emerson, geologist of the Louisiana station and in charge of soil survey work in the State, died October 11. 74 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. A. E. Grantham, head of the department of agronomy in Delaware College and agronomist of the Delaware station for the past twelve years, has resigned, effective February I, to become manager of the agricultural service bureau of the Virginia-Carolina Chemical Com- pany, with headquarters at Richmond, Va. Thomas Jahne and R. R. Mulvey are now assistants in soils in the college of agriculture of Purdue University. Ove F. Jensen has resigned his position with the department of agronomy, Iowa State College, and on December 8 became midwest agronomist for the Soil Improvement Committee of the National Fertilizer Association, with headquarters in Chicago. Bradford Knapp, for the past several years in charge of extension work in the Southern States for the United States Department of Agriculture, has resigned effective January 15 to become dean and director of the Arkansas college and station, succeeding Martin Nelson. Theodore E. Odland has been made assistant in agronomy and J. E. Chapman has been elected assistant in soils in the Minnesota col- lege of agriculture. George A. Pond is now assistant professor of farm management in the Minnesota college of agriculture. Robert R. Smith has been elected agronomist of the Northwest Ex- periment Farm, Crookston, Minn. J. L. Snyder, president of the Michigan Agricultural College from 1896 to 191 6, died during the last week of October at his home in Michigan. He was born in Butler Co., Pennsylvania, in 1859 and was a graduate of Westminster College. He was president of the Michi- gan Agricultural College during the period of its greatest prosperity and growth. Dr. Simon Fraser Tolmie is the new Canadian Minister of Agri- culture. Dr. J. H. Grisdale, director of the Central Experiment Farms and acting deputy minister, has been made deputy minister of agriculture, and has been succeeded as director by E. S. Archibald, formerly animal husbandman at the Central Experiment Farms. F. L. Wagner of the Bureau of Plant Industry has succeeded George S. Knapp as superintendent of the Garden City, Kans., sub- station, Mr. Knapp having resigned to become State irrigation com- missioner. AGRONOMIC AFFAIRS. 75 C. B. Waldron, dean of the North Dakota college of agriculture, is on leave for a year to act as director of agricultural education in the United States army. F. S. Wilkins, who has been assistant in agronomy at the Oregon college since July, has returned to Iowa State College to take charge of forage crop experimentation, succeeding O. F. Jensen. E. L. Worthen has resigned as associate professor of agronomy in Pennsylvania State College and has been appointed extension profes- sor of soil technology for 1919-20 at Cornell University, vice E. O. Fippen, who is on a year's leave of absence. A National Farm Crops Improvement Association was formed on December 2, 1919, at the Stock Yards Inn, Chicago, during the Inter- national Live Stock and Grain Show. The object of the organization is to promote the work of the various State crop improvement and ex- perimental associations. Meetings will probably be held annually in connection with the International Live Stock and Grain Show. Prof. R. A. Moore of Wisconsin is president ; Prof. G. H. Cutler of Alberta, Prof. Manley Champlin of South Dakota, and Prof. John Buchanan of Iowa, vice-presidents; and Prof. J. W. Nicholson of Michigan, secretary-treasurer. The topics discussed at the meeting were: "What Should Constitute Pedigree or Purebred Seeds," by C. P. Bull of Minnesota ; " Seed Inspection and Certification by Associa- tions," by J. W. Nicholson of Michigan and B. S. Wilson of Kansas; " Official State Seed Inspection," by A. L. Stone of Wisconsin ; and "Marketing Pedigree Seeds," by H. D. Hughes of Iowa. Each paper was followed by discussion. Aid for National Research Council. The National Research Council has issued the following state- ment which will be of interest to members of the American Society of Agronomy, as our Society has representation on the Council. The Carnegie Corporation of New York has announced its purpose to give $5,000,000 for the use of the National Academy of Sciences and the National Research Council. It is understood that a portion of the money will be used to erect in Washington a home of suitable architectural dignity for the two bene- ficiary organizations. The remainder will be placed in the hands of the Acad- emy, which enjoys a Federal charter, to be used as a permanent endowme.pt for the National Research Council. This impressive gift is a fitting supplement to Mr. Carnegie's great contributions to science and industry. The Council is a democratic organization based upon some forty of the great scientific and engineering societies of the country, which elect delegates to it? 76 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. constituent divisions. It is not supported or controlled by the Government, dif- fering in this respect from other similar organizations established since the be ginning of the war in England, Italy, Japan, Canada, and Australia. It intends if possible to achieve in a democracy and by democratic methods the great scientific results which the Germans achieved by autocratic methods in an autocracy, and at the same time to avoid the obnoxious features of the auto- cratic regime. The Council was organized in 1916 as a measure of National preparedness and its efforts during the war were mostly confined to assisting the Government in the solution of pressing war-time problems involving scientific investigation. Reorganized since the war on a peace-time footing, it is now attempting to stimulate and promote scientific research in agriculture, medicine, and industry, and in every field of pure science. The war afforded a convincing demonstra- tion of the dependence of modern nations upon scientific achievement, and nothing is more certain than that the United States will ultimately fall behind in its competition with the other great peoples of the world unless there be persistent and energetic effort expended to foster scientific discovery. Journal of the American Society of Agronomy. Plate 3. Fig. i. Plants of a continuously self-fertilized strain, showing uniformity in structural details, evenness in position of ear, and height of plant. Fig. 2. First-generation crossed plants, showing that they retain the uniformity of the parent strains and are exceedingly vigorous. Journal of the American Society of Agronomy. Plate 4. Fig. i. The uniform excellence of crosses between inbred strains, due to the fact that barrenness, degeneracy, and poor heredity in general have been largely eliminated, counts heavily in maximum production. Fig. 2. The identical plants shown above, lined up for closer ' inspection. With one exception, the ears were placed at the sixth node, an indication of their remarkable trueness to form. Journal of the American Society of Agronomy. Plate 5. Fig. i. Ears of a variety, two inbred strains derived from it after ten gen- erations of self-fertilization, and the first-generation hybrid between these two strains, showing proportional yields in adjoining rows. The actual yields (left to right) were 96, 32, 20, and 115 bushels to the acre. Fig. 2. Ears from inbred strains (left) and of crosses between them. This illustrates, from actual field results, the method of combining inbred strains to obtain maximum yields. Journal of the American Society of Agronomy. Plate 6 Fig. i. The three rows in the foreground are the first, second, and third generations of a cross between the two inbred strains next to the left of the largest row. Fig. 2. Plants from the rows shown in Figure I, in the same order, showing that self-fertilization after crossing again reduces the plants to the level of vigor of their inbred parents. Journal of the American Society of Agronomy. Plate 7. £ r e r r d G C Fig. i. Seeds from inbred strains (top and bottom) and from the first generation of a cross between them (middle). The first-generation hybrid between inbred strains is handicapped by its poor start, having to grow from small seeds borne on reduced inbred plants. The method of double crossing overcomes this advantage and plants starting from large, well-developed seeds like those shown in the center are capable of greater production. Fig. 2. Ears of a double cross which are nearly equal in size and are uni- form, considering the great germinal diversity of such a complex hybrid. The specimens here shown were produced by pollinating a first-generation cross between two white strains with a cross of two yellow strains, as shown in Plate 5, figure 2. Similar results can be obtained with all yellow or all white strains, thus avoiding the mixture of colors on the ears. JOURNAL OF THE American Society of Agronomy Vol. 12. March, 1920. No. 3 SELECTION IN SELF-FERTILIZED LINES AS THE BASIS FOR CORN IMPROVEMENT. 1 D. F. Jones. Introduction. The way in which domesticated animals and cultivated plants have been brought to their present condition is, for the most part, un- written history. In the chief crops and most valuable animals the changes which have taken place during their amelioration have been vast. Selection, following up more or less intentional and uninten- tional hybridization, is the agency mainly responsible for these great advances. Selection has been practiced in different ways. At the start primi- tive man chose forms from the wild which suited his need or fancy. This selection involved nothing more than the appearance of the indi- viduals chosen. Selection based solely on appearance remained the only method until recent times and is still largely in vogue today with many plants, particularly Indian corn or maize, the most valua- ble plant in the Western Hemisphere. Animal breeders as a rule have been more progressive than plant improvers. In the eighteenth century, when the great breeds of cat- tle, horses, and swine began to assume their present familiar forms and patterns, there was a far-reaching change in the method of selec- tion. This innovation was the pedigree record system. The essen- tial purpose of a pedigree system is to make possible selection based upon performance. When the characters and capacity of the ances- tors and of the progeny as well as of the individuals themselves were 1 Contribution from the Connecticut Agricultural Experiment Station, New Haven, Conn. Received for publication November 28, 1919. 77 yS JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. taken into consideration, progress was immediately made more cer- tain, judging from the quick rise to popularity of the Shorthorns, Herefords, Percherons, Southdowns, Berkshires, and other familiar breeds. The names of Blakewell, Colling, Bates, Booth, Tompkins, Price, and Hewer are well known and stand for skill in applying systems of mating based primarily upon ancestry and not appearance alone. To them, the ability to beget offspring with the desired char- acters and vigorous growth was of first importance. When one considers whether or not these principles which have worked such marvels in the animal realm can be applied to plants he is faced with a different situation. Hermaphroditism, which is the rule with plants, introduces a complicating feature. However, many plants, of which corn is a good example, are continually and almost universally cross-fertilized naturally, so that for all practical purposes such plants may be looked upon as bisexual organisms. The obstacles in the way of applying a pedigree record system, as used with animals, directly to the corn plant are so great as to render it wholly impracticable. To carry over the method unaltered one would have to select individual plants in the field before flowering time to be used as females and as males. Such plants, would need to be artificially pollinated. The resulting progenies should then be grown in order to determine the relative value of the individual par- ents which were selected for mating. This would involve too great labor. What makes such a method still more hopeless is the fact that the chief character aimed at, production of grain, is not visible until after fertilization is accomplished. Therefore, it is not to be wondered at that the pedigree record system as applied in animal breeding has not seemed to offer much help* to the plant worker. The choosing of seed corn still proceeds in the primitive way of se- lection based solely upon appearance. The choice can be on only one side of the family, for no matter how excellent an ear of corn may be there is no way of judging the qualities of the plants which fur- nished the pollen to fertilize the seeds on that ear. Biometry tells us, however, that a sample of the hundred or so individuals representing the plants which supplied the pollen, taken at random from the field, will come close to the average of the whole population. In other words, one can be sure that the finest ear in any open-pollinated field of corn will have mediocre parentage on the staminate side. As a matter of fact it will be worse than mediocre because a random sam- ple of pollen from a corn field does not of necessity come from the average producers of grain. Many plants which produce poor yields JONES : SELECTION IN SELF-FERTILIZED LINES. 79 supply pollen abundantly. Some plants in every field are wholly barren in pistillate parts but lack nothing in their staminate function. In general, there is a negative correlation in the development of the two inflorescences in corn. The situation is exactly the same as if the animal breeder would turn his choicest dams loose upon the open range to be served by any sire happening to come along, a foolhardy procedure which is never tolerated. Yet the same method in corn breeding receives sanction and encouragement because no better method of selection has so far been available. By propounding elaborate score cards for judg- ing seed corn, agronomists have thrown a cloak of pseudo-science over an antiquated system which is anything but scientific. Perhaps this is too severe an arraignment in "view of the natural limitations imposed. In any case, there are abundant data from the work of Biggar (i), 2 Hartley (n), Hutcheson and Wolfe (14), Love (18), Love and Wentz (19), McCall and Wheeler (20), Montgomery (21), and Olson, Bull, and Hayes (24) to show that the correlation be- tween the appearance of the seed ear, other than mere size, and the performance of the progeny is negligible. To be sure, opinion has not been unanimous. Where correlations have been found, however, they are low and, when significant, have nearly always involved size relations where some association is to be expected. The same factors which enable the mother plant to produce a large ear tend to make the progeny large also. No one can reasonably urge that selection in field-pollinated corn is not worth while. Who would be so foolish as to recommend the planting of the poorest ears to be found in a field? The picking of the best- looking seed ears has undoubtedly had everything to do with the building up the varieties as they now are. But too close attention to details defeats the end chiefly sought — maximum production. Corn breeders all along have recognized the great disadvantage which the unknown parentage on half of the family tree entails, and have endeavored to establish certain systems of selection based upon performance which would make possible a taking of the best from this unknown material. The various means of applying the ear-to- row method of breeding have been steps in the right direction, but they have never gone very far for the reason that selection is made after fertilization has taken place at random, whereas to hold out hope of real success judgment must be passed before the individuals are mated. Furthermore, the ear-to-row method has brought in a 2 Reference is to " Literature cited," p. 98. SO JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. complicating feature. In every case there has been a narrowing of the network of related individuals which results in a loss of vigor. The good which might be accomplished by close selection is offset by reduction in size and rapidity of growth due to inbreeding. What is needed at present is a different viewpoint, a radical change in method, and an application of sound principles, already proved in animal breeding experience, adapted to the peculiar conditions which hermaphroditism involves and which will obviate the dis- astrous effects of inbreeding. A new system is desired which is workable, which will make improvements permanent, and which will hold out hope for continued betterment. Before outlining what the writer believes is a beginning of such a system, it is necessary to sum up briefly present theories in regard to the significance of inbreeding and the results of actual experiments involving relationship matings. Significance of Inbreeding. The fact of reduced size and lessened vigor resulting from consan- guinity, particularly in domesticated animals and plants which have previously undergone long periods of outcrossing, is beyond question. Due to Darwin's theories, this loss of growth ability has been looked upon as a natural phenomenon arising from an assumed physiological necessity for germinal mixing. The Knight-Darwin dictum that " nature abhors perpetual self-fertilization," like many popular slo- gans, has carried conviction but obscured the real point at issue. The practical stockmen have been closer to the true meaning of in- breeding than the biologists. The former have held that the effects of inbreeding, whether good or bad, were due to a concentration of like characters, favorable or unfavorable as the case might be. The key to the solution of the problem was withheld until Mendel's theory of heredity became known. Segregation of recessive factors was seen to account for much of the evil brought to light by inbreed- ing, but segregation of independent Mendelian units was inadequate to explain the universality of the phenomenon. Recombination in a definite proportion of cases would bring all the favorable growth fac- tors together in the pure breeding or homozygous condition, which would then be unaffected by any system of mating. As this was clearly not the case with many organisms, corn, for example, always being reduced by self-fertilization, it remained for the great exten- sion of the knowledge of heredity which has taken place in the past ten years to make possible a logical interpretation of the whole problem. JONES : SELECTION IN SELF-FERTTLTZED LINES. 8i The development of the chromosome theory of heredity, initiated by Bateson and Punriett in England in their studies on coupling of factors and carried forward with such remarkable success by Mor- gan and his school in this country, has shown conclusively that heredi- tary factors are carried in groups and that it is these groups of factors which mendelize. Because independent recombination within the groups does not occur, it is rarely possible to get all the more favor- able characters collocated in one individual. Crossing, therefore, brings together the greatest number of different factors. As favor- able growth characters tend to be expressed rather than unfavorable ones, whenever the two are paired, the complementary action of dominant factors furnishes a logical means of understanding the beneficial effects of crossing and the reverse results of close mating. For the first time, inbreeding is viewed in a clear light. The results of this system of mating depend solely upon the inheritance received. Consanguinity in itself is in no way injurious. Like the detective who unearths a crime, it should be praised rather than cen- sured. The important consideration is the constitution of the stock before close mating is practiced. It is therefore not strange that some species should be naturally self -fertilized. Such forms came to possess fortunate combinations of all, or nearly all, the most favorable factors for development. Cross-fertilization could add nothing and was dispensed with. One should not confuse hybrid vigor with the important part that hybridi- zation has had in evolution. Sexual reproduction certainly serves a very useful purpose or it would never have been established as the dominant method of propagation in nearly all forms of life. Ex- ogamy, making possible a greater elasticity in adaptiveness to new conditions, is the primary importance of sexual reproduction. Hy- brid vigor is a secondary result, however, of considerable value. Results of Self-Fertilization with Corn. Alt'ho the evidence is convincing that consanguinity is not inher- ently harmful, one should not blind himself to the fact that the imme- diate results of close breeding in particular instances are very injurious or even disastrous. Take the case of corn. The fact that every time ordinary varieties of this crop have been self-fertilized a diminution in growth has taken place is established by the investi- gations of East (7, 8), East and Hayes (9), East and Jones (10), Shull (26, 27, 28, 29), Hayes (12, 13), Montgomery (22), Noll (23), and Jones (15). The reduction in size, however, proceeds to a cer- 82 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. tain point and then stops. Beyond that point there is no further change. The stage at which the alterations cease corresponds to the point where change in visible characters and reduction in variability also stop. In other words, a heterozygous complex is changed to a number of homozygous strains. Their form, size, characteristics, and ability to grow depend on the inheritance handed out to them as a chance allotment subject to the automatic elimination of the most unfavorable characters. During this change to constancy, particularly in the earlier gen- erations of self-fertilization, a great many characters appear which are clearly unfavorable to growth. Some of them cause the imme- diate extinction of the individual possessing them. To cite a few examples, there are several forms of chlorophyll deficiency more or less complete as indicated by pure white, virescent, and yellow seed- lings. Other stages of chlorophyll reduction permit growth, but at a lessened rate, for example, japonica striping, green striping, fine striping, and golden chlorophyll color. Dwarf plants are commonly found in inbred strains of corn as well as many forms of partial or complete sterility of one or the other inflorescence or both. All these represent defective germ plasm. They may be expected in every field of corn and are quite commonly found, but in small quantities, due to the fact that constant crossing keeps them hidden from sight. Differences in height, in breadth and color of leaves, in size and shape of ears, and, in fact, in every detail of the plant structure are also segregated into different lines, so that when constancy is reached each strain is characterized by certain features which make it distinct from every other strain coming from the same original variety. As these types isolated by inbreeding differ in external characters, so also do they differ in ability to grow. Theoretically, all possible levels of vigor may result from inbreeding. Plants may be isolated in the first generation which are incapable of reproduction. Other plants may theoretically be found which show no reduction in vigor at all. So far, no such plants have been discovered in ordinary varieties of corn, but they may be hoped for. The expectation can be diagrammed as shown in figure 4. The solid lines represent strains which have already been obtained; dotted lines show what may be found later. There may even be some lines which show an increased size and vigor resulting from self-fertilization. This fol- lows from the fact that the factors combined in the heterozygous condition in ordinary corn probably do not show perfect dominance, so that when all the most favorable factors are combined together in JONES : SELECTION IN SELF-FERTILIZED LINES. 83 the diploid state a heightened growth efficiency may be expected. The chance for obtaining such strains in corn is exceedingly small, as will be shown later, but the possibility holds out hope for the maxi- mum improvement of this plant. 12345 676 9 10 GENERATIONS Fig. 4. Graph showing diagrammatically the actual and theoretical results of self-fertilizing ordinary varieties of corn. The solid lines represent strains already obtained which have either become extinct or have been reduced to 50 percent or less of the vigor of the original variety. The broken lines represent strains which can be expected theoretically when this plant is worked with more extensively. Darwin (6) obtained self-fertilized morning-glories which were larger and finer than the original variety with which he started. Miss King (17) has secured inbred rats which are at least no less vigorous than the stock at the start. The greater uniformity of inbred strains of corn is one of their most pronounced features. Table 1, in which the coefficients of variability of the original variety and several self-fertilized strains derived from it are given, shows how uniformity is obtained in meas- urable characters. This feature is even more noticeable in the details of the plants which are difficult of statistical expression, as seen in the accompanying illustrations (Plates 3 to 5). 84 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Table i. — Variability of a variety of com, of several inbred strains derived from it after ten generations of self-fertilisation, and of crosses between these strains. 0. Pedigree No. Coefficients of variability. Height of plant. Length of ear. Number of nodes. Rows of grain. Ori p"ina1 \tc\ ri^r v 8.8l 19.07 10.07 14.22 1-6-1-3 4-95 18.16 5-66 7-88 1-7-1-1 8.29 19.52 7.41 8.74 1-7-1-2 8.00 19.69 7-35 9.23 1-9-1-2 6-39 n-55 5.89 7.65 (1-6-1-3 X 1-7-1-1) Fi . 6.42 19-39 5-73 8.75 (1-6-1-3 X 1-7-1-2) Fi 495 14.72 4.99 8.09 (1-6-1-3 X 1-9-1-2) Fi 5-59 13.09 5-33 7-94 (1-7-1-1 X 1-9-1-2) Fi 8.01 16.79 6.15 10.24 (1-7-1-2 X 1-9-1-2) Fi 6.49 19.29 5-89 9.60 a These figures have been compiled from Tables 4, 5, 6, 7, 18, 19, 20, and 21 of Conn. Agr. Expt. St'a. Bui. 207. The way in which the approach to uniformity and constancy pro- ceeds depends upon the system of mating practiced. It is most rapid in self-fertilization for the reason that when a plant becomes homo- zygous for any character it must ever after remain so. Brother and sister or parent and offspring matings are much less efficient for the reason that a homozygote may mate with a heterozygote. Thus, six generations of self-fertilization are more effective in producing homo- zygosity than seventeen generations of brother and sister mating. The usual way in which an inbreeding experiment with corn is conducted is to select individual plants, self-pollinate them, and use one in each generation as the progenitor of the line. In that case, the rate of approach to homozygosity depends upon the constitution of the individuals chosen. Theoretically, a plant may be completely homo- zygous or completely heterozygous in every generation. The only thing which must follow is that no individual can be more hetero- zygous than its parent ; it may be the same or less. On the average, however, there is a reduction of half the number of heterozygous allelomorphic pairs in each generation in self-fertilization. The rea- son for this may be made clearer by figure 5. As an illustration, an organism heterozygous for 15 units is self- fertilized. The progeny of such an individual will be distributed in respect to the number of heterozygous and homozygous chromosome pairs according to the curve which resembles the familiar probability curve. The bulk of the individuals come at the center in the median grades of complexity. Therefore, an individual selected at random JONES : SELECTION IN SELF-FERTILIZED LINES. 85 to be the progenitor of the next generation, as a rule, will be half as heterozygous as its parent, and so on in each generation until com- plete homozygosity is ultimately obtained. Representing the homo- zygosity at the start as o percent, after seven generations of self- fertilization it will be 99 percent on the average and, after twelve gen- erations, 99.9 percent. This is the theoretical expectation with inde- pendent inheritance. Linkage and differential viability enter as com- plicating factors. Actual results with corn show a very high degree of stability after six generations of self-fertilization and after ten generations no appreciable changes have been apparent. HETEROZYGOUS PAIRS 15 14 1? 12 11 10 9 8 7 6 5 4 3 2 1 HOMOZYGOUS PAIRS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fig. 5. Graph showing the theoretical distribution of the progeny of a self- fertilized organism heterozygous in fifteen allelomorphic pairs. The bulk of the individuals are in the central area. Any individual selected at random as a progenitor of a self-fertilized line will therefore most likely be about half as heterozygous as its parent. This will continue to be the case until homozygosity is obtained. Value of Inbred Strains. The resultant plants which have been continuously self-fertilized until uniformity and constancy are reached have gone through a process of purification whereby much defective germplasm has been eliminated. Sterility is one of the first things eliminated. All other \ 86 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. markedly injurious characters soon fall by the wayside, due to the handicap they put upon the individuals possessing them. The result is that the surviving plants possess much of the best that was in the variety to start with, whether conscious selection has been practiced or not. Inbreeding removes the support which hybrid vigor sup- plies and each character must pass the rigid scrutiny of natural selec- tion on its own merits. The result is beneficial. When inbred strains are crossed, vigor is at once regained and the uniformity of the parent strains is retained in the first generation of the cross. This is a very valuable feature which cannot be overemphasized. Not all combinations give equal results, but when a desirable cross is made its performance is very gratifying. Picture a field of corn in which each plant is like every other plant and in which there are no barren plants, no abnormals or degenerates, where every member of the population is contributing an equal share towards the production of grain. Other things being equal and given a perfect stand of plants, a greater yield can be looked for from such a field than is possible by any other method of corn breeding now known. It does not take such a very large ear of corn to weigh a pound. Most ears of dent corn on exhibition in corn shows weigh somewhat more than this. It can easily be figured that an acre of corn with a perfect stand which produces even a half-pound ear on every plant will yield close to 100 bushels of grain. All plants are never exactly alike in practice, as not all plants have an equal oppor- tunity to grow. Unfavorable situation, accidental injury, or disease may reduce the production on some plants. Given an equal oppor- tunity to grow, however, with crosses of inbred strains the same results are obtained because all the plants are alike in hereditary constitution. Many crosses at the Connecticut Agricultural Experi- ment Station have yielded more than 100 bushels to the acre, out- yielding all ordinary varieties grown on the same field. Another point of great importance is that, with due allowance for seasonal variation, the same result can be obtained every time the particular cross is made. In other words, any improvement that is made is permanent because the parental strains are constant and give the same results every time they are crossed. Other secondary ad- vantages are secured. If a certain cross is found good for a par- ticular district, seed to produce that corn can be grown almost any- where that corn can be successfully raised. For example, a cross giving excellent results in Illinois could be made in Pennsylvania or California and the seed continually produced there provided the JONES: SELECTION IN SELF- FERTILIZED LINES. 87 plants were properly grown and matured. If change in locality has any effect it is beneficial, as Collins' (3) results on new-place effect indicate. A further advantage lies in the fact that it gives the origi- nator of valuable strains of corn the same commercial right that an inventor receives from a patented article. Selection in Self-Fertilized Lines. As inbreeding has been largely looked upon until now, it is a method of purifying stock automatically by freeing it of defective germplasm. There is no reason why it should stop there. What is most desired is a means of obtaining the very best that is in the mate- rial in hand. This is the aim of intelligent selection. Self-fertiliza- tion, by making possible a reliable estimate of the hereditary values of the male as well as the female parent, furnishes the best and surest means of gaining this desired result. Looked at in this way, the inbreeding and crossing of inbred strains as such is without particular value. As a system of breeding, however, selection in self-fertilized lines offers results in proportion to how extensively and skillfully itjsjused. It is to be expected that not all crosses between inbred strains will give better results than the original variety. As a matter of course the bulk of the germplasm in every variety is commonplace. Of necessity, most inbred strains will have medium value. As will be explained later, the first-generation cross between inbred strains is handicapped at the outset of its growth. A means is at hand to over- come this. Even if many strains are found which give no appre- ciable increase in yield over the original variety, this does not vitiate the value of this method of selection, if sometime and somehow superb germplasm can be isolated which will give greatly increased yields. It then remains to compare the cost of producing the seed in the new way with the increase in yield obtained to decide whether or not the method is practicable. Selection in self-fertilized lines makes possible the only thoro means of carrying on breeding with cross-fertilized plants, based upon performance and not appearance of the seed parent alone. The application of this system to corn is exceedingly simple, altho it involves considerable tedious work. A tentative plan of procedure may be outlined as follows. From a number of plants of the variety to be worked with, individual plants are selected in the field for vigorous growth shortly before flowering and artificially self-polli- 88 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. nated. 3 It should again be emphasized that the chance for real improvement depends primarily on the extent of the material in- cluded. Let us take 100 self-pollinated ears to start with. About 20 percent of failures of hand pollination will occur, due to various causes, so that more than 100 plants will have to be pollinated arti- ficially. We have found that two men can make about a thousand hand pollinations in a season of a month and a half. If many more than ioo plants of one variety are to be worked with it is necessary to plant at different times to extend the flowering period. Starting with ioo self-pollinated ears of corn, each is to be planted the following year in a separate plat and again certain plants chosen for hand pollination. Five pollinations is the minimum number to make in each plat to insure a fair prospect of continuing the line safely. As grain production can never be properly judged before maturity, it is well to make as many pollinations as possible and select the plants to continue the line at the end of the season. Plant as many self ed ears in each line the following year as possible. Let us say three ears in each line are grown. As the plants grow in the field, it is possible in most cases to determine before flowering which of the three progenies is the best, so that only one of the three is chosen to be hand-pollinated to continue the line. From this, three hand-pollinated ears are again selected at maturity and the process is repeated until uniformity and constancy are reached. This can be expected ordinarily in about six generations, altho variations will be found. Selecting for the most vigorous plants will tend to per- petuate hybrid plants, as these give the larger growth. Much re- mains to be determined as to the best method of carrying on selection in this way. The procedure outlined may be diagrammed as shown in figure 6. A modification of this method at the start has been suggested by Mr. B. H. Duddleston 4 of the U. S. Department of Agriculture, whereby a large number of naturally pollinated ears are selected and planted in an ear-to-row test, saving part of the seed of each ear. The remaining seed of those ears which gave the best results are planted the following year, certain plants hand pollinated, and selec- tion carried on in some such way as outlined above. By this method one year is lost at the start, but the chances of including the best material in the plants at the beginning is greater. 3 For rapid and careful work in pollination the writer has found nothing better than ordinary paper bags of the best quality, using 8-pound bags for the ears and io-pound or 12-pound bags for the tassels. 4 In a letter. JONES : SELECTION IN SELF-FERTILIZED LINES. 89 In the course of such a period of intense inbreeding there will be great loss of size and vigor. Many strains will become extinct be- cause of sterility, dichogamy, or extreme weakness. Many such strains could be maintained by sib mating, but it seems better to adhere rigidly to self-fertilization alone. Any other procedure is merely putting off the day of judgment which all plants must pass sooner or later. Many strains will possess undesirable characters which will make it seem best to discard them. Here is a serious LINE A LINE B LINE C etc. I I I PLANTS of : Fig. 6. Diagram showing the method of selection in self-fertilized lines. A number of plants of an ordinary variety are self-fertilized and each becomes the starting point for an inbred line. From the progeny of each certain plants are again self-fertilized and their progenies grown, but only one is chosen to continue the line. All the individuals in every generation in any one line trace back directly to the original progenitor. problem which needs much more investigation. In general, there is a correlation between the productiveness of inbred strains and the crosses derived from them. This correlation is far from perfect, however, and some extremely poor strains give astonishingly fine results when crossed in certain combinations. Until we are more familiar with selecting in inbred material it seems wiser to keep all strains which are able to make fair growth until their value can bt tested in combination. The following are some of the most important qualities, in the opinion of the writer, for which to select : gO JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Normal chlorophyll development generally indicated by dark green color with absence of striping. Stocky plants with short internodes and well developed brace roots which stand perfectly upright thruout the season. Bright colored, plump seeds. Freedom from mold on the ear. Freedom from parasitism on the plant or ear, particularly smut. Absence of suckers. Many other characters are important, but these named are sug- gestive. One of the most gratifying features of crosses between in- bred strains is their freedom from mold on the ear. Inbred strains have been unconsciously highly selected for this behavior because ears enclosed in bags are very apt to become moldy. Many specimens are lost completely for this reason and of course cannot be perpetuated. Some strains are also highly resistant to smut damage, as previously noted (16) and as shown in Table 2. This is a very valuable fea- ture. Collins (4) also shows that corn can be made somewhat im- mune to ear-worm damage. It is possible that something can be done to reduce the damage from other serious pests of corn. Table 2. — Susceptibility to smut infection shown by four inbred strains of corn, by the original variety from which they were derived thru self-fertilisation for ten generations, and by first generation crosses between some of these strains. 1917. 1918. 1919. Total. Strain No. No. of Smut in- No. of Smut in- No. of Smut in- No. of Smut in- plants. fection. plants. fection. plants. fection. plants. fection. Per- Per- Per- Per- cent. cent. cent. cent. 992 1,000 1. 00 144 1-39 2,136 0.56 Inbred strain 1-9-1-2 596 •34 559 • 71 157 1.91 1,212 .69 Inbred strain 1— 7-1-2 408 .49 307 9.12 145 4.14 860 4.19 950 9-79 599 25.87 198 8-59 1.747 15-17 (1-6-1-3 X 1-9-1-2) Fi. . . 3i 3i (1-6-1-3 X 1-7-1-2) Fi . . . 435 3.9H 3.14 4.346 2.83 (1-6-1-3 X 1-7-1-1) Fi. . . 439 2.28 326 .61 765 i-57 Original variety 114 1-75 250 .40 119 7-56 483 2.48 Utilization of Inbred Strains. After some degree of uniformity has been attained and there is no further drop in vigor it is necessary to test the surviving strains in all possible combinations with each other. After preliminary crosses have been made by hand pollination and tested in a limited way, the best of these can then be tried on a more extensive scale. The easiest way to do this is to use one strain as a pollen parent and plant all the JONES : SELECTION IN SELF-FERTILIZED LINES. 9 1 different lines in the same crossing plat, detasseling all but the one kind. If several such plats sufficiently well isolated from other corn are available and a different strain is used as pollinator in each, the testing can be done in a few years and the best combinations iden- tified. As a general rule, reciprocal crosses will give about the same result. This is not always true, however, and this fact should be taken into consideration when making final choice. When two inbred strains are combined, the first-generation hybrid plants are handicapped because of the small seeds from which they start, these having been grown on reduced, nonvigorous, inbred plants. Considerations are involved other than merely the amount of food material stored in the seed. Whatever the cause may be, it is appar- ent that the hybrid seedlings start off slower than seedlings of varie- ties or even second-generation seedlings grown from large seeds produced on F x hybrid plants, altho the F 1 surpasses the F 2 at the end of the season (15). This handicap is greater in some seasons than in others. Moreover, as most inbred strains are low in yield, the cost of first-generation hybrid seed to be used for general field planting is very great as long as inbred strains are used to produce such seed. A method which overcomes both these objections is found in a cross of two first-generation hybrids of such genetic constitution that heterozygosity is not reduced and hybrid vigor is retained at the maximum. Such doubly crossed plants, starting from large, well- developed seeds grown on vigorous first-generation hybrid plants, are free from the handicap at the beginning of growth and for this reason yield more than the first crosses. Such a double cross must be made by combining four inbred strains which are all genetically unlike. They may all come from the same original variety or from different varieties. The essential point is that they should give good results when crossed singly in the several possible combinations. Even then it is not certain that the double cross will give the yields desired. This must be established by trial. A valuable combination once obtained, however, can always be had every time the same strains are united in the proper way. When two homozygous strains are brought together all the plants of the first-generation hybrid are exactly alike hereditarily. W^hen two dissimilar F 1 crosses are again combined the result is a population in which nearly every plant is genetically dif- ferent, to a greater or less extent, from every other plant. As re- gards development, each plant is essentially a first-generation hybrid because every chromosome pair is composed of unlike elements, the 92 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. parental strains presumably differing in all chromosomes. Practi- cally no recombinations can take place which allow like factors to come together and weaknesses to appear. There is, however, great germinal diversity, so that the beautiful uniformity of a first-genera- tion hybrid is sacrificed. It is largely the evenness in structural de- tails which is lost ; in ear formation and production of grain double crosses show a remarkable ability to yield. Ears used as parent strains in making a double cross, together with ears from the result- ing hybrid, are shown in Plate 5, figure 2. Ears produced by such a double cross are shown in Plate 7, figure 2. Plants from the parent strains and the first-generation hybrids are shown in Plate 6, and kernels in Plate 7, figure I. Some idea of the chromosome situation in such a complex hybrid may be had from an illustration. If a first-generation cross between two self-fertilized strains which is heterozygous in three chromo- some pairs, 'for example, as A A', B B', C C, is crossed with another single hybrid having chromosomes of different constitution, such as A" A'", B" B'", C" C", then the double cross will have various com- binations of these chromosomes as: A A"', B' B", CC", or A' A", B B'", C C", etc. If there is no crossing over there will be 64 dif- ferent combinations, but each association of chromosomes will be a heterozygous one. No like chromosomes are paired, therefore all plants are essentially first-generation hybrids. With ten chromo- somes together, as in maize, with breaks in the linkage in unknown amount, the total number of possible types is exceedingly great. The way in which such a double cross is put together is illustrated in the following diagram : Variety Inbred strains Single crosses Double crosses The inbred parental strains when reduced to homozygosity after 6 to 10 generations of self-fertilization are constant and permanent. The double cross is not and if self-fertilized will go back to about the same level of vigor as the selfed material started with. If the cross is allowed to pollinate at will, in a few years it would constitute an JONES I SELECTION IN SELF-FERTILIZED LINES. 93 ordinary variety of about the same characteristics as the variety originally used. In all this combining of strains one should not lose sight of the main principle involved. The crossing has no potency in itself. It is merely a short-cut method giving the best factors contained in the parental strains an opportunity of expressing themselves in comple- mentary action. The value of the double cross lies solely in the good fortune which has attended the selection of the plants while they were being brought to homozygosity. It is there that effort and skill should be extended. Thoroness is also needed in discovering the best plan of bringing the strains together. Certain incidental advantages make this method of double crossing more practicable than single crossing. First, the crossed seed for general field planting is produced abundantly and therefore cheaply. For example, inbred strains that yield from 20 to 40 bushels produce around 100 bushels per acre when crossed. As the seed is to be grown in a crossing plat where every alternate row is detasseled and the product of the detasseled row only used for planting, the double crossing method gives from 40 to 50 bushels of seed per acre, while the older method furnishes only 10 to 20, a difference which may easily determine feasibility. A second great advantage lies in the fact that incomplete detasseling of all the first-generation seed mother plants is not so serious as when inbred strains are used. As applied commercially it will no doubt be difficult to get 100 percent cross- fertilization in a crossing plat owing to some of the plants not being detasseled at the proper time. Plants of the F 1 generation give second-generation plants when pollinated with their own kind of pollen which are not markedly inferior to plants of the first genera- tion. Selfed strains pollinated with their own pollen, instead of crossed as they should be, give inbred plants which are very poor in yield. Even a moderate number of such plants in a field of corn might nullify all the advantages to be obtained by crossing. On the other hand, the complexity of the new method is a serious handicap. Four strains instead of two must be maintained. How- ever, inbred strains are easily carried on after they have been brought to homozygosity. As all plants are exactly alike, it makes no dif- ference whether they are cross-pollinated or self-pollinated within the strain. They can therefore be grown in an isolated plot and no extraordinary pains are necessary to exclude foreign pollen, for whenever outcrossing has taken place the resulting plants can be easily spotted the next year because of their greatly increased size 94 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. and can be removed before pollen is shed. In view of the great gain to be had it does not seem possible that the disadvantage of com- plexity will long stand in the way of the general utilization of this method of selection. Carrier (2) has called attention to the increase in weight of seed immediately resulting from cross-fertilization as a method of increas- ing yield. The phenomenon of heterosis in both the endosperm and embryo has been well established. The stimulation is manifest in greater weight, higher specific gravity, and a more rapid rate of maturing. However, this effect is merely a by-product of hybrid vigor. If it is worth while to have heterosis working in the seed it is many times more desirable to have that stimulus in the plant, as it is there that yielding capacity is largely determined. As a matter of fact, increase in weight of seed due to cross-pollination is an indi- cation that the plant is not working at its maximum. Mixed pollina- tion between different inbred strains giving reciprocally crossed and selfed seeds on the same ears have shown increases in weight of crossed seeds of as high as 35 percent. Similar mixed pollinations made between two first-generation hybrids where the plants them- selves were several times more vigorous than inbred strains have given much smaller increments in weight, ranging from 1 to 9 per- cent. A double cross, which has given the highest yield so far ob- tained, when pollinated with a mixture of its own and pollen from a very distinct source, gave no appreciable difference in weight of crossed seed. Cross-pollination enters as a disturbing factor in varietal trials, but the effect works both ways and each variety has an equal chance to respond to the stimulus of hybrid vigor. However, cross-pollination tends to increase more the yields of the poorer yielding varieties and for that reason varietal tests are somewhat misleading. On the other hand, it is difficult to see how the tests can be conducted so as to avoid this, for, unless varieties are grown close together, differences in soil will more than offset any differences due to unequal effects of cross-pollination. Comparative Cost of Crossed Seed. Naturally, hybridized seed must sell for a higher price than ordi- nary seed. Corn is America's greatest crop. It produces more total value than any other plant grown in this country. It also produces a fairly large value per unit of area. Along with this the cost of seed in relation to the crop produced is far lower than that of most cultivated plants, as shown in Table 3. JONES: SELECTION IN SELF-FERTILIZED LINES. 95 Table 3. — Comparison of cost of seed of some farm crops in relation to the returns. - Crop. Quantity of seed Market price per bushel. Cost of seed Average yield per acre. Value of crop per acre. v^ost 01 seed as percent- age of returns. Bushels. Cents. Bushels. Potatoes 10. 48.9 54-89 IO9.5 $53-56 9-13 Wheat i-5 98.6 I.48 16.6 16.41 9.02 Barlev 2.0 54-3 I.09 25.8 14.00 7-79 Oats 2.0 43-8 .88 29.7 12.99 6.77 Rice 2.0 92.4 1.85 34-i 31-50 5.87 Beans •5 600.0 3-oo 15.0 90.00 3-33 Corn •3 64.4 .19 25.8 16.65 1. 14 The figures for the quantity of seed planted) per acre are taken from Bailey's Cyclopedia of American Agriculture and those for market' value and yield per acre are for the year 1914 from the United States Department of Agricultural Yearbook for 1914. A fair comparison on the basis of cost of seed is difficult because prices for all seed stocks fluctuate greatly. They are generally higher than the market price at which the crop is sold. At the same time, the market value is an average figure and furnishes a means of com- parison, altho the cost of seed corn is usually somewhat more in proportion to market value than is the case with the other crops mentioned. Is not the farmer fully justified, then, in trebling or quadrupling his outlay for seed corn if he can be reasonably certain of a 10- percent increase in yield as well as an improvement in quality due to a lessened amount of moldy grain? A crossing plat a half acre in size should furnish abundant seed for 50 acres. If the material to plant were furnished, the amount of time and effort expended on the crossing plat would be very little greater than that needed for searching a 50-acre field of corn for seed ears, as now practiced. A gratifying feature of such hybridized seed is that all the grain that is fully developed and properly matured can be used for planting. As all seeds are equivalent in germinal constitution, the scrubbiest ear in the lot is the equal of the handsomest. This is a novel situation that will require proof to convince most corn growers. Possibility of Obtaining True Varieties of Corn. The proposed method of selecting in self-fertilized lines and cross- ing in double combinations undoubtedly holds the greatest hope for the improvement of corn at the present time. The utilization of hybrid vigor, however, is a makeshift measure. It would be desir- able to dispense with crossing altogether if possible. Theoretically, if all the factors contributed by the parental strains to make the hybrid valuable could be gathered together in one plant, that plant 96 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. would be the homozygous progenitor of a variety of corn which would be as stable as any naturally self-fertilized species, such as wheat, peas, beans, tomatoes, tobacco, etc. In fact, for the first time there would be a true variety of corn. So-called varieties of corn at the present time are merely germinal hodge-podges. What would be of greatest value is that such a variety would have no less vigor than the hybrid. Greater growth and productiveness are expected from such a homozygous type because dominance is seldom perfect. Duplex combinations are more favorable for maximum developmental energy than simplex relations, provided the same desired factors are all pres- ent in the former as in the latter. For example, let us assume two inbred strains, each of which con- tributes four factors chiefly responsible for the increased vigor when these two strains are crossed. One has the diploid composition AA, CC, EE, GG; the other BB, DD, FF, HH. The hybrid is haploid in regard to each factor pair, viz., Aa, Bb, Cc, Dd, Ee, Ff , Gg, Hh. The fact that more different dominant factors are here present working together is the basis for interpreting the phenomenon of hybrid vigor. If these eight factors can be recombined into a homozygous instead of a heterozygous union which would be AA, BB, CC, DD, EE, FF, GG, HH, such a type should be appreciably more efficient in its life processes even than the hybrid. Such an organism would behave the same whether crossed or self-pollinated. In corn it would be a form hitherto unattained except possibly in one instance, but we may hope that such forms may be obtained. At least, every effort should be put forth in this direction. Investigational work along this line has great opportunities. However, let us not deceive ourselves as to the magnitude of the task ahead. As yet no one can estimate the number of factor differ- ences in corn concerned with growth vigor. Fifty definite hereditary units have been positively identified and more are being found. Many of these play no important part in development. Perhaps the number of factors really necessary for vital processes is not as great as now imagined. With only one factor in each chromosome group, there would be ten independent units, making necessary the growing of 1,048,576 plants in order to have an even chance of getting the one plant recombining all these factors in a homozygous state. With more than one factor in each chromosome, as there unquestionably is, the numbers necessary to work with are stupendous. Two fac- tors in each chromosome so spaced as to have 10 percent breaks in the linkage with each other would necessitate 20 20 individuals in the JONES : SELECTION IN SELF-FERTILIZED LINES. 97 segregating generation to have an even chance of securing the one plant desired. This number of corn plants would require an area roughly 3,700,000,000,000 times the area of the United States. Even if it were possible to grow this number, how could this one plant be identified so as to be protected from cross-pollination? The immediate attainment of such homozygous plants with maxi- mum vigor from ordinary varieties of corn is utterly impossible at the present time. More refined methods of genetic analysis are needed. Further knowledge of the factor relations in maize in re- gard to the location in the chromosomes is of the utmost value. A hopeful sign is furnished by a variety of corn obtained from the southwestern Indians which has been self -fertilized by Collins for three generations and so far has shown no reduction in vigor. It is from just such isolated communities where the corn has probably not been outcrossed for long periods of time that such a state of affairs is to be expected. Natural selection has eliminated all but the most favorable characters. On account of isolation this corn has approached the condition of naturally cross-fertilized species which show no decrease in artificial self-pollination. What Nature has done with time as an ally we may be able to do by better methods of genetical analysis than are now available. Table 4. — Differences in number of totally barren plants as shown by the fre- quency distributions of test plats of corn. a Source. Number of barren plants per plot. 012 34567 10 Ave. Num- number ber of of plats. barren plants. 30 3-03 24 1.04 54 .63 50 .76 Percent- age of plants, barren. Varieties Inbred strains Single first-generation crosses Double first-generation crosses a About 120 plants in each plat 2-53 .87 •53 .63 As continued self-fertilization automatically eliminates much unde- sirable heredity, it seems perfectly feasible to use inbreeding as a process of purification. After a number of homozygous strains of corn have been obtained they can be crossed among themselves and a new variety re-created with justified expectation that a real improve- ment can be effected. Such a variety would not be much different from the old one and would be but little more stable. With a great 5 The number of combinations is calculated from the formula r+i) n_1 ] 2c where r -f- 1 is the linkage ratio, in this case 10 percent or 9 -f- 1, n is the num- ber of factors in each chromosome, and c is the number of chromosome pairs. 98 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. deal of sterility, chlorophylll deficiency, and abnormal tendencies re- moved, however, the variety ought to be somewhat better able to grow and to produce. Table 4 shows how total sterility is removed by inbreeding. The results obtained will depend very largely on how extensively selection is carried on while the plants are being self- fertilized and after crossing. This is almost an untouched field for improvement and is equally applicable to all naturally cross-fertilized plants and bisexual animals. Conclusions. It is desired that thruout all this discussion of proposed methods for improvement the fundamental principle will not be lost to sight. This basic tenet is that selection in self-fertilized lines makes possible a reliable estimation of heredity values of both sexes. This has never heretofore been accomplished with corn. When selection is made on as extensive a scale as the prospects for improvement justify, there is reason to believe that inbred strains will be produced which are much better than any that have so far been obtained. If it is possible to obtain self-fertilized lines only slightly less productive than ordinary varieties it may then be desirable to dispense with the method of double crossing outlined above and use only single crosses. Vigorous self ed strains when crossed should give practically as great yields as any double cross and would be more uniform and much easier to produce. The data collected at the Con- necticut station indicate clearly that for the present the combination of two first-generation hybrids is necessary 'to secure maximum yields. When, therefore, a method which is both commercially remunera- tive and scientifically exact is available, are the agronomists of this country going to be slow in applying it ? The important need at the present time is extensive investigation aiming to make the system as workable as possible. Such investigation merits the intelligent inter- est and active support of all corn growers, seed dealers, and agrono- mists. This interest has not been lacking and it is to be hoped that corn improvement is now entering upon a new era. Literature Cited. 1. Biggar, H. Howard. The relation of certain ear characters to yield in corn. In Jour. Amer. Soc. Agron., v. 11, no. 6, p. 230^234. 1919. 2. Carrier, Lyman. A reason for the contradictory results in corn experi- ments. In Jour. Amer. Soc. Agron., v. 11, no. 3, p. 106-113. 1919. 3. Collins, G. N. New-place effect in maize. In Jour. Agr. Research, v. 12, no. 5, p. 231-243. 1918. JONES : SELECTION IN SELF-FERTILIZED LINES. 99 4. . Tropical varieties of maize. In Jour. Heredity, v. 9, no. 4, P- 147" 154. 1918. 5. Cunningham, C. C. The relation of ear characters to yield. In Jour. Amer. Soc. Agron., v. 8, no. 3, P- 188-196. 1916. 6. Darwin, Charles. The Effects of Cross and Self-Fertilization in the Vegetable Kingdom, viii -j- 482. London : Appleton & Co. 1875. 7. East, E. M. Inbreeding in corn. In Conn. Agr. Expt. Sta. Rpt. for 1907, p. 419-428. 1908. 8. . The distinction between development and heredity in inbreeding. In Amer. Nat., 43: 173-181. 1909. 9. , and Hayes, H. K. Heterozygosis in evolution and in plant breeding. U. S. Dept. Agr., Bur. Plant Indus. Bui. 243, 58 p. 1912. 10. , and Jones, D. F. Inbreeding and Outbreeding : Their Genetical and Sociological Significance, 277 p. ' Philadelphia : Lippincott Co. 1919. 11. Hartley, C. P. Progress in methods of producing higher yielding strains of corn. In U. S. Dept. Agr. Yearbook for 1909, p. 309. 1910. 12. Hayes, H. K. Normal self-fertilization in corn. In Jour. Amer. Soc. Agron., v. 10, no. 3, P- 123-126. 1918. 13. ■ , and East, E. M. Improvement in corn. Conn. Agr. Expt. Sta. Bui. 168. 1911. 14. Hutcheson, T. B., and Wolfe, T. K. Relation between yield and ear characters in corn. In Jour. Amer. Soc. Agron., v. 10, no. 6, p. 250-255. 1918. 15. Jones, D. F. The effects of inbreeding and crossbreeding upon develop- ment. Conn. Agr. Expt. Sta. Bui. 207. 1918. 16. . Segregation of susceptibility to parasitism in maize. In Amer. Jour. Botany, v. 5, no. 6, p. 295-300. 1918. 17. King, H. D. Studies on inbreeding. I. The effects of inbreeding on the growth and variability in body weight of the albino rat. In Jour. Expt. Zool., v. 26, no. 1, p. 1-54. 1918. 18. Love, H. H. The relation of certain ear characters to yield in corn. In Proc. Amer. Breeders' Asso., 7: 29-40. 191 1. 19. , and Wentz, J. B. Correlations between ear characters and yield in corn. In Jour. Amer. Soc. Agron., v. 9, no. 7, p. 315-322. 1917. 20. McCall, A. G., and Wheeler, Clark S. Ear characters not correlated with yield in corn. In Jour. Amer. Soc. Agron., v. 5, no. 2, p. 117, 118. 1913. 21. Montgomery, E. G. Experiments with corn. Nebr. Agr. Expt. Sta. Bui. 112. 1909. 22. . Preliminary report on effect of close and broad breeding on pro- ductiveness in maize. Nebr. Agr. Expt. Sta. 25th Ann. Rpt., p. 181-192. 1912. 23. Noll, C. F. Experiments with corn. Pa. Agr. Expt. Sta. Bui. 139. 1916. 24. Olsen, P. J., Bull, C. P., and Hayes, H. K. Ear type selection and yield in corn. Minn. Agr. Expt. Sta. Bui. 174. 1918. 25. Pearl, Raymond, and Surface, Frank M. Experiments in breeding sweet corn. In Ann. Rept. Maine Agr. Expt. Sta., p. 249-307. 1910. 26. Shull, G. H. The composition of a field of maize. In Proc. Amer. Breeders' Asso., 4: 296-301. 1908. IOO JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. 27. . A pure line method of corn breeding, In Proc. Amer. Breeders' Asso., 5: 51-59- I909- 28. . Hybridization methods in corn breeding. In Amer. Breeders' Mag., 1 : 98-107. 1910. 29. . The genotypes of maize. In Amer. Nat., 45: 234-252. 191 1. THE RELATION OF SIZE, SHAPE, AND NUMBER OF REPLI- CATIONS OF PLATS TO PROBABLE ERROR IN FIELD EXPERIMENTATION. 1 James W. Day. The data for the studies discussed in this paper were obtained at the Shelbina field of the Missouri Agricultural Experiment Station. The plat used was approximately one-fourth of an acre in extent and was apparently very uniform thruout. It was situated on the Putnam silt loam. In the fall of .1916, Fulcaster wheat was drilled at the rate of 5 pecks per acre in 100 rows 8 inches apart and 155 feet in length. The following June the wheat was harvested by hand in 5-foot row segments, and the yield of grain and of straw was recorded for each unit. There were, therefore, 3,100 units available for study. A calculation was first made to determine the direction in which the greatest variation in yield existed. To indicate the variation that existed in the direction of the rows, the yields of the thirty-one series of 100 adjacent 5-foot row segments were compared. To indicate the variation across the rows, the yields of plats composed of three adjacent 155-foot rows were studied. The units of comparison in the two instances were of practically the same area. It was found that there was much greater variation across the rows than along them, or, in other words, the rows extended in the direction of least variation. THE RELATION OF SIZE OF PLAT TO VARIATION IN YIELD. A study was made of the relation of the size of plat to variation in yield. The standard deviation and coefficient of variability were 1 Contribution from the Oklahoma Agricultural and Mechanical College, Stillwater, Okla. Received for publication December 28, 1919. The writer gratefully acknowledges the invaluable advice of Prof. C. B. Hutchinson, now professor of plant breeding at Cornell University, in plan- ning the investigation discussed in this paper. day: probable error in field experiments. ioi determined for each of many sizes of units that were formed by com- bining 5-foot row segments. The results are presented in Table I. Table i. — Data obtained on relation of size of plat to variation in yield. Coeffi- Coeffi- Size of plat. Standard cient of Size of plat. Standard cient of deviation. varia- deviation. varia- bility. bility. 3 adjacent 155-ft. rows 5.885.40 14.19 5 adjacent 50-ft. rows 3,227.66 14.49 5 adjacent 155-ft. rows 9,022.00 13-07 10 adjacent 50-ft. rows 5.443.76 12.13 10 adjacent 155-ft. rows 13,018.31 9-43 15 adjacent 50-ft. rows 6,802.67 I0.I8 20 adjacent 155-ft. rows 19,405.81 7-03 20 adjacent 50-ft. rows 7,4I7-98 8.32 1 5-ft. row 164.84 37.20 30 adjacent 50-ft. rows 7,295.68 5-46 1 10-ft. row 262.89 29.58 50 adjacent 50-ft. rows I5,I35-7I 6.79 1 1 5-ft. row 354-07 26.52 3 adjacent 15-ft. rows 749-30 18.68 434-67 24.41 5 adjacent 15-ft. rows 1,101.14 16.49 518.07 22.81 10 adjacent 15-ft. rows 1,694.70 12.72 1 30-ft. row 602.27 22.53 20 adjacent 15-ft. rows 2,659-38 9.98 1 35-ft. row 661.48 21.32 50 adjacent 15-ft. rows 4,960.03 7-45 728.73 20.28 100 adjacent 15-ft. rows 3.688.95 2-77 796.60 19.85 100 adjacent 30-ft. rows 6,326.87 2.38 1 50-ft. row 922.68 20.67 8 adjacent 5-ft. rows 595-02 16.77 1 60-ft. row 1,021.62 18.99 16 adjacent 5-ft. rows 869.78 12.36 1 75-ft. row 1.309-56 19.64 24 adjacent 5 -ft. rows 1,120.85 IO.54 1 100-ft. row 1,502.29 16.74 32 adjacent 5-ft. rows 1,238.05 8.74 1,912.73 17.01 64 adjacent 5-ft. rows 1,599.08 5-66 1 150-ft. row 2,333-92 17-36 96 adjacent 5-ft. rows 1,786.46 4.20 3 adjacent 50-ft. rows. 2,186.24 16.37 100 adjacent 5-ft. rows i,787-98 4.02 Table I shows the effect of increasing the size of the plat and bears out the accepted theory that in general an increase in the size of the plat increases the accuracy of the results obtained. It is evident from the data presented here that an increase of the size of the plat up to at least one twentieth of an acre will decrease variation, but that minor exceptions to this general truth occur in the case of very narrow long plats that run in the direction of least 'variation. It should be noted also that single plats of the maximum size used in this study, one twentieth of an acre, ordinarily do not reduce the variation to a point that will permit the detection of differences be- tween varieties or fertilizer treatments. THE RELATION OF SHAPE OF PLAT TO VARIATION IN YIELD. Data on the relation of shape of plat to variation in yield are given in Table 2. The data in Table 2 show that the shape of the plat has an impor- tant effect on the accuracy of experimental results. In each of the divisions of the table, plats of approximately the same area but of different shapes have been taken. The results show conclusively 102 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Table 2. — Data on relation of shape of plat to variation in yield. Units of w comp No. of adjacent rows. iich plat is osed. Length of rows. Length of rows in plat. Shape of plat. Standard deviation. Coefficient of varia- bility. X* 661. Feet. I 150 150 Long in direction 01 least 2.333-92 17-30 variation 3 5° jso uu. 2 186 24 IO -37 1 I 5 15° Rectangular t f\r\ a *7 r\ x ,uy4- /V 12.72 24 5 120 Long in direction of most 1,120.85 10.54 variation 5 155 775 Long in direction 01 least 9,022.00 13-07 variation 15 50 750 ao. AO A 6,802.67 10.18 5° I 5 75° -L^UIlg III Kill CL.LXU11 Ul IllUoL 7-45 variation 100 5 500 do. 1,787.98 4.02 10 155 I .55° .Long m direction 01 least 13,015.31 9-43 variation 30 50 1,500 Somewhat long in direction 7,295.68 5.46 of least variation 100 15 1,500 Long in direction of most 3,688.95 2-77 variation 3 155 465 Long in direction of least 5,885.40 14.19 variation 20 15 300 Square 2,659.38 9.98 1 5° 5° t • a- n Long in direction 01 least 68 922. 20.7 variation 3 15 45 do. 749.30 18.68 8 5 40 Square 595-02 16.76 5 50 250 Long in direction of least 3,227.66 14.49 variation 10 15 150 do. 1,694.70 12.72 32 5 160 Long in direction of most 1,238.05 8-74 variation that those plats having their greatest dimensions in the direction of least variation are more variable than plats that approximate squares in shape ; and squares in turn are more variable than plats having their greatest dimension in the direction of greatest variation. That the more uniform results obtained in the latter plats are due entirely to the shape of the plat is borne out by the fact that in no case are these plats larger than the corresponding ones that lie lengthwise in the opposite direction. The results indicate that where single plats of a given area are to be used, the greatest accuracy can be obtained from long narrow plats lying in the direction of greatest variation. In an experimental area that was as variable in its length as in its width, the shape of day: probable error in field experiments. 103 the plat would exert no influence. When an investigator is unable to ascertain in which direction his soil is most variable, the use of square plats is probably advisable. If long narrow plats are used in such a case, the chances are even that they will extend in the direc- tion of least variation and, therefore, be less reliable than square plats. A better method, however, consists in ascertaining the nature of the soil in a preliminary test and using long narrow plats that have their greatest dimension in the direction of greatest variation. The data thus far presented indicate that it is possible to obtain fairly accurate results from single plats under certain conditions. On the area which forms the basis of these studies, however, no plat that was much below one twentieth of an acre in area or that had its greatest dimension in the direction of least variation gave a coeffi- cient of variability of less than 2.5 percent. A series of plats each composed of 100 thirty-foot rows (see Table 1) gave a coefficient of variability of 2.38 percent. This variation is sufficiently low to indi- cate that a plat of such size and shape would give accurate results. The plat just mentioned was slightly less than one twentieth of an acre in area and extended 66.6 feet in the direction of greatest varia- tion. It is probable that on an experimental area where the plats could be so arranged that the length in the direction of greatest varia- tion could be increased much beyond 66.6 feet, while the plat width was decreased, a plat of even smaller area would give fairly depend- able results. THE RELATION OF REPLICATION OF PLATS TO VARIATION IN YIELD. Data on the relation of replication of plats to variation in yield are presented in Table 3. The effect of using a unit of comparison composed of systematically distributed parts is shown in Table 3. A comparison of the data in the first division of that table with those in Table 1 makes strikingly evident the fact that a unit composed of replicated rows gives more accurate results than a continuous unit of the same area. In every instance there was much less variation in units composed of repli- cated rows than in units of similar size composed of adjacent rows. A comparison of the data in Tables 1 and 3 also justifies the con- clusion that the same degree of accuracy can be obtained on a much smaller area when a unit composed of replicated parts is used. A study of the left half of Table 3 shows furthermore that as the num- ber of replications of rows of a given length is increased, more accu- rate results are obtained, as also is the case when the number of 104 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. replications remains the same and the size of the parts replicated is increased. Table 3. — Data on relation of replication of plats to variation in yield. Composition of row unit. Standard deviation. Coefficient of variability. Composition of block unit. ■ Standard deviation. Coeffi- cient of bility. No. of rows. Length of rows. No. of blocks. JNo. ot adja- cent rows. Length of rows. Feet. Feet. 3 155 3.972.75 95« 5 3 50 2.054.38 3-97 5 155 3.663.36 5-31 10 3 50 4,481.34 3-35 10 155 5,992.95 4-34 14 3 50 6.549.04 3-50 15 155 5,820.50 2.0 I 5 5 50 3,277.14 2-94 20 155 7.835-5° O Si A 2.04 10 5 50 5.309.54 2.30 3 50 1,590.41 T T Q A II. oO 7 5 15 2-343-53 5-04 5 50 it 657-22 7-43 14 5 15 3,057.07 3-3 2 10 50 2,651.13 5-95 28 5 15 3.395-89 I.O3 15 50 2,934-72 4-39 3 10 15 2,385.96 5-97 20 50 3.939-99 4.42 7 10 15 -> a 8A 08 3,4°0.2o 3-75 30 50 4.55°-9i 3-40 14 10 15 3,000.94 I. OO 60 So 5.497-99 2.06 3 20 x 5 3,584-34 4-5 1 3 15 T 1 A1 6 20 15 2,Ol6.34 I 27 6 15 923-25 11.52 3 50 15 10,566.53 5-29 12 15 1,217.31 7-59 5 8 5 1,334-90 7-53 18 15 1,542-75 6.42 10 8 5 1,541-28 4-35 26 15 1,428.38 3-97 52 8 5 2,077-35 I-I3 52 15 2,197-54 3-05 5 16 5 1,530.27 4.32 78 15 3,177-47 2.89 7 16 5 1,936.41 3-90 104 IS 3.383-64 2-35 9 16 5 2,132.85 3-34 156 IS 3,984-77 1.85 14 16 5 2,857.67 2.88 18 16 5 3,123.77 2-45 36 16 5 1,679-44 0.66 5 100 5 4.456.79 2.00 5 3 155 10,183.75 4.92 In the right half of Table 3 are shown the data obtained by repli- cating blocks that contain three or more rows or parts of rows. This material proves that in general the conditions that apply to the repli- cation of rows are also applicable to the replication of blocks. Repli- cated blocks invariably gave more accurate results than the same area in a single block. The same degree of accuracy was obtained on a smaller area when replication was employed. As the number of replications of blocks of a given area was increased, more accurate results were obtained. Where the number of replications remained the same, an increase in the size of the block replicated decreased the coefficient of variability. The data also show that the best shape of block for replication is that which gave the best results with single plats, that is, a long nar- row plat with its greatest dimension in the direction of greatest varia- tion. Five replications of three adjacent 155-foot rows, for example, day: probable error in field experiments. 105 gave a coefficient of variability of 4.92 percent, while five replications of a plat 5 feet in width and running entirely across the experimental area gave a coefficient of variability of 2.00 percent. The areas of the two series were approximately the same, but in the first case the greatest dimension of the individual blocks was in the direction of least variation, while in the second case it was in the direction of the greatest variation. CONCLUSIONS. From the results obtained, it is evident that : 1. Increasing the size of the plat to at least one twentieth of an acre, and probably much beyond, reduces variation. 2. The shape of the plat has an important effect on variation that has in the past been overlooked or misunderstood. More accurate results are obtained from single plats that are long and narrow and extend in the direction of greatest variation than from those of other shapes. Square plats or approximately square plats are to be pre- ferred to long narrow plats that have their greatest dimensions in the direction of least variation. 3. The results from single plats are usually not sufficiently accurate to determine small differences between varieties or between fertilizer and cultural treatments. 4. The use of a unit of comparison composed of systematically distributed parts gives results that are much less variable than those obtained from an equal area in a single plat. 5. An increase in the number of replications of a plat of given size increases the accuracy of the results. 6. When the number of replications remains constant but the size of the plat replicated is increased, variation is reduced. 7. The most effective replicated block from the point of view of shape is one that is long and narrow and has its greatest dimension in the direction of greatest variation. I06 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. THE USE OF A SELECTION COEFFICIENT. 1 C. H. Myers. INTRODUCTION. In connection with the conduct of some corn-breeding work, 2 the writer has found a " selection coefficient " to be of considerable use in helping to make selections. A report of the use and application of this coefficient may perhaps be of service to others engaged in similar work. PURPOSE OF THE SELECTION. The selection was begun for the purpose of producing strains of dent corn better adapted to New York conditions. As the number of silos increased there was an increased demand for dent corns. Seed of these for the most part was obtained from the Central States, where the season is more suitable for corn growing. As a result of this practice, much of the dent corn produced in New York did not reach a sufficiently advanced stage of maturity, even for the best ensilage purposes. The question of maturity was of prime importance. It was real- ized, however, that if maturity alone was considered it could be at- tained, of course, but probably at the expense of yield, which was also important. It was desirable to consider both of these qualities in conducting the selection work. METHOD OF DETERMINING THE COEFFICIENT. An adaptation of the individual ear-to-row method was followed. It is not the purpose here to describe the details, further than to say that 100 ears were chosen for the beginning and that each year of the experiment each ear in the plat was replicated at least once. The plat furnishing the material for the basis of this report was located in Saratoga Co., N. Y. 3 The elevation at this place is 400 feet. The 1 Paper No. 80, Department of Plant Breeding, Cornell University, Ithaca, N. Y. Received for publication January 10, 1920. 2 This work was begun by Dr. H. J. Webber in 1908 and conducted by him until 1912. Since that time it has been in charge of the writer. Credit is due to Dr. H. H. Love for first suggesting the use of this coefficient. 3 This work was done in cooperation with Mr. G. R. Schauber, Ballston Lake, N. Y., to whom much credit is due for the successful conduct of the experiment. MYERS! USE OF SELECTION COEFFICIENT. I07 growing season has an average length of 154 days, with the average date of the last spring frost falling on May 6 and the average date of the first fall frost falling on October 7. 4 Only the early flint corns could be matured here in normal seasons. This location was, there- fore, an excellent one for a corn plat planned for the purpose outlined above. The yielding capacity of the different rows was determined at har- vest time by weighing the corn produced in each one. These weights were recorded as " total yield per row." Another index of the yield- ing capacity was obtained by dividing the total weight of corn in each row by the total number of eared stalks in that row. This gave the " average yield per stalk." This latter method is possibly a more accurate measure than the former, especially in cases where the stand is not perfect. It is recognized that the latter is not an absolute measure, for the yield per stalk is very likely to be heavier in case of a thin stand. In the experiment under consideration, however, only ears of good germination were used as seed and the planting was made at a thicker rate than desired, the plants being thinned later to the proper stand. This method resulted in rather a uniform stand, altho there was always some slight variation in the number of stalks per row. The maturity of each individual row was determined as follows : All the ears were sorted into two lots, ripe and unripe. The two lots of ears were then counted and the number of ripe ears was divided by the total number of ears, to give the expression which is called " percentage of maturity." This method of sorting into ripe and unripe is more or less an arbitrary one and is not absolute. On the average, however, if the work is done carefully, it results in a definite enough measure of maturity. Having determined the yield and the maturity, it was desirable to combine both of these expressions into a single one which should serve as a basis for selection. This was done by multiplying the average yield per stalk by the percentage of maturity. This gave the " selection coefficient " which was used. Using the total yield per row instead of the average yield per stalk gave another coefficient. This was not markedly different from the first, which was the one adopted. T;ible 1 shows a portion of the data obtained from the Saratoga County plat in 1909. This table illustrates not only the method of obtaining the coefficients but also the amount of indi- 4 Wilson, Wilfred M. Frosts in New York. N. Y. (Cornell Univ.) Agr. Expt. Sta. Bui. 316. 1912. 108 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Table i. — Data from the first 50 rows, Saratoga County plat, crop of 1909, from which the selection coefficients were obtained. Row No. Progeny No. Number of stalks. Number of suckers. Total yield. Yield per stalk. Number ripe. Number unripe. Per- centage turity. Selection coefficient (total yield X percent- turity). Selection coefficient (yield per stalk X percent- turity). Lbs. Lbs. Pet. Pet. I 3- 9-1 73 3 50.5 0.692 19 53 26.4 13-33 O.183 2 3-i 5-1 56 16 42.5 • 759 19 30 38.8 16.49 .294 3 3-21-1 74 2 48.0 .649 30 40 42.9 20.59 .278 4 3-22-1 7i 3 5i-5 •725 18 47 27.7 14.27 .201 5 3-27-2 75 3 44.0 .586 21 44 32.3 14.21 .I89 6 3-3 1-1 74 5 53-5 .723 II 60 15-5 8.29 .112 7 3-32-1 75 2 51.0 .680 19 52 26.8 13-67 .182 8 3-33-1 73 3 43-5 .596 24 40 37-5 I6.3I .223 9 3-34-1 7i 3 50.0 •704 13 57 18.6 9-30 .131 10 3-35-1 74 2 47-0 .635 23 43 34-9 16.40 .222 11 3-37-1 72 2 44.0 .611 17 51 25.0 11.00 •153 12 3-40-1 74 2 48.0 .649 25 43 36.8 17.66 •239 13 3-42-2 75 3 45-o .600 13 53 19.7 8.87 .118 14 3-43-1 72 6 47.0 •653 37 30 55-3 25-99 .361 15 3-46-1 7i 2 39-5 •556 14 36 28.0 11.06 .156 16 3-47-1 69 5 52.0 •754 20 50 28.6 14.87 .216 17 3-49-1 73 5 51.0 .699 18 50 26.5 13-52 .185 18 3-64-1 75 3 42.5 •567 21 42 33-3 14-15 .189 19 3-61-1 72 4 50.0 .694 27 40 40.3 20.15 .280 20 3-60-1 72 4 39-5 • 549 27 30 47-4 18.72 .260 21 3-50-1 68 8 46.0 .676 4 55 6.8 3.12 .O46 22 3-67-1 73 2 36.0 •493 13 36 26.5 9-54 .131 23 3-75-3 70 4 41.5 •593 16 4i 28.1 11.66 .167 24 3-80-1 74 7 45-5 .615 29 34 46.0 20.93 .283 25 3-90-1 75 46.0 .613 23 46 33-4 I5-36 .205 26 3-94-2 70 2 42.5 .607 24 38 38.8 16.49 .236 27 3-95-1 70 4 49.0 .700 7 59 i.i -52 .008 28 3- 9-2 72 9 50.5 .701 29 37 43-9 22.17 .308 29 3-21-2 72 5 44-5 .618 19 46 29.2 12.99 .l80 30 3-22-2 68 8 47-5 .699 42 26 61.8 29-36 •432 3i 3-2 7-3 66 4 44.0 .667 16 47 25-4 11. 18 .169 32 3-31-2 74 5 52.5 .709 41 30 57-8 30.35 .410 33 3-32-2 73 9 5i-5 • 705 10 57 14.9 7.67 .105 34 3-33-2 62 4 39-5 .637 26 24 52.0 20.54 •331 35 3-34-2 72 2 48.5 .674 30 39 43-5 21.10 •293 36 3-35-2 73 3 49.0 .671 20 44 31-2 15.29 .209 37 3-37-2 70 6 43-5 .621 22 36 37-9 16.49 •235 38 3-40-2 69 6 46.0 .667 19 40 32.2 14.81 .215 39 3-41-1 72 5 47-5 .660 31 32 49.2 23-37 .325 40 3-42-4 68 8 46.5 .684 13 49 21.0 9-77 .144 41 3-43-2 73 5 5i-0 .699 59 15 79-7 40.65 •557 42 3-46-2 73 6 48.5 .664 50 19 72.5 35-i6 .481 43 3-47-2 73 5 44-5 .610 35 36 49-3 21.94 .301 44 3-49-2 75 4 4i-5 •553 30 29 50.9 21.12 .281 45 3-50-2 73 4 46.0 .630 37 32 53-6 24.66 , .338 46 3-60-2 75 5 49-5 .660 34 37 47-9 23-71 .316 47 3-61-2 72 10 49-5 .688 5i 20 71.8 35-54 .494 48 3-64-2 7i 5 49-5 .697 44 30 59-5 29-45 .415 49 3-67-2 74 5 49.0 .662 20 50 28.6 14.01 .189 50 3-90-3 73 4 35-5 .486 39 16 70.9 25.17 •345 MYERS : USE OF SELECTION COEFFICIENT. IO9 vidual variation among the different rows with respect to both yield and maturity. For this year the yield of individual rows ranged from 30.5 pounds to 55.0 pounds, while the percentage of maturity ranged from 1. 1 to 87.3. The value of the selection coefficient ranged from 0.8 to 62.2. The rather close agreement between the two coefficients is shown graphically in figure 7. The solid line represents the coefficient ob- tained by using the total yield per row, while the broken line repre- sents the one obtained by using the average yield per stalk. These two lines follow each other very closely. USE OF COEFFICIENTS IN MAKING SELECTIONS. At harvest time a number of ears from each progeny were reserved as seed ears. Wired tags were used for labeling. This procedure was followed because it was not feasible to make the calculations for the selection coefficient at the time of harvesting. It was easier to choose from six to ten seed ears from each progeny and reserve them, if necessary, until planting time, as this did not require much time or storage space. After the calculations described above were made, selections of the best rows for that year were made on the basis of the selection coefficient, the greater coefficients indicating the better rows from the standpoint of yield and maturity. The reserved seed ears from the unselected rows could then be discarded. As stated above, each ear in the plat was always replicated at least once. * It is interesting to compare the coefficient of selection for the same ear planted in different rows. It should be borne in mind that the arrangement of planting was such that these rows were not side by side but in different parts of the plat. A comparison of these two series of coefficients for the different years shows that they follow each other rather closely. That is, a row which has a high coefficient in one series has a correspondingly high one in the other series. This is illustrated graphically in figures 8 and 9, taken from data obtained in 1909 and 1910. The solid line in each case represents the coeffi- cient for the first series, while the broken line represents it for the second series. The agreement here is not absolute, due partly at least to soil inequality, but in a general way they follow very closely, especially in the case of the extremes. RESULTS OF THE SELECTION. The remnants of the original seed ears with which the plat was started were saved and at different times during the course of the HO JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. MYERS : USE OF SELECTION COEFFICIENT. I 1 I experiment served to furnish seed for a comparative test as to the progress of the selection. In each of these comparative tests, from five to ten rows 50 hills long, planted from a mixture of the original seed ears, were alternated with the same number of rows planted from a mixture of seed ears from the selected rows of that year. For convenience the former is labeled " original " and the latter " se- lected " in the table displaying these results. Table 2 shows the results from the comparison plat in 191 1, after three years of selection. The difference as shown by the yield per stalk is not striking, the average for the selected being 0.621 pound while that for the original is 0.650 pound. The average yield per stalk of the original is heavier than that of the selected, due to the immaturity of the former. When we compare the percentage of maturity, there is a striking difference, the average for the selected being 71.9 while that for the original is 13.2. In the last column of the table are given the selection coefficients for these rows, to illustrate the value of this in making selections from such material. Table 2. — Data from the comparison plat in Saratoga County in 1911 and 1912. Data from 191 i. Selected seed. a Original seed. b Row No. Yield per Percentage Selection Row No. Yield per Percentage Selection stalk. of maturity. coefficient. stalk. of maturity. coefficient. II O.673 66.7 0.449 12 0.643 21.6 O.I39 13 .656 73-2 .480 14 .708 4.8 •034 15 .690 77-5 •535 16 .738 12.8 .094 17 .640 73-2 .468 18 .677 9-7 .066 19 .690 77-3 •533 20 .670 13-2 .088 na .649 65-2 .423 I2a •655 29.7 .194 13a .500 71.4 .400 14a .633 17.8 •113 15a .490 69.2 •339 16a .602 6.7 .040 17a •565 67-5 .381 18a .583 10.8 .063 19a .600 77-5 .465 20a •594 5-i .030 Data from 1912. 1 O.609 4i-5 0.253 2 0.776 3 •554 56.7 .314 .800 5 .729 62.9 .458 t •875 7 .650 63.8 .415 8 •740 9 .706 67.2 •474 10 •743 5-4 .040 Average. . .650 58.4 ' .383 ' .787 1.1 .008 The selected seed consisted of remnants of ears selected from the 1910 crop for 191 1 planting. & The original seed was a mixture from the remnants of the original seed from Illinois. I 1 2 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. In 1 91 2 another such comparison p'lat was grown. The results from this plat are also summarized in Table 2. Again there is no striking difference between the selected and original with respect to the average yield per stalk. The latter has a somewhat heavier yield per stalk, but this again is due to its immature state. The per- centage of maturity for both is greater in 191 1 than in 1912, on ac- count of an unusually early fall frost in the latter year. The per- centage of maturity for the selected is 58.4 as compared with 1.1 for the original. Here again the selection coefficient has been calculated to illustrate its application. The individual ear-to-row selection resulted in the isolation of five progenies of the original parent ears. Since 191 3 no ear-to-row selections have been grown. Only mass selection from the standing corn in the field is now practiced. Varietal tests in which this strain is included indicate that it maintains its characteristc of early ma- turity combined with good yield. BAHIA GRASS. 1 John M. Scott. A large number of species of Paspalum are native to Florida. These are generally known as blanket grass or water grass, but some- times as goose grass. In addition, several valuable species have been introduced from South America, including Dallis grass (Paspalum dilatatum) , Vasey grass (Paspalum larranyagai) , and lastly the spe- cies here discussed, Paspalum notatum, native in South America and northward to Mexico, for which the Bureau of Plant Industry sug- gests the name of Bahia grass. It gives most promise as a pasture grass. This was introduced into the United States in 1913 by the Bureau of Plant Industry under S. P. I. No. 35067. 2 Another intro- duction, S. P. I. No. 37996, was made in 1914. Bahia grass was first planted at the Florida Agricultural Experi- ment Station in May, 1913. The original plat is still growing. From the very first this grass gave promise of being valuable. On March 31, 191 5, a plat of Bahia grass was planted in the pas- ture on the experiment station farm. The 'ground was plowed and a good seed bed prepared. Plants were taken from the original seed 1 Contribution from the Florida Agricultural Experiment Station, Gainesville, Fla. Received for publication March 5, 1920. 2 Accession number of the Office of Foreign Seed and Plant Introduction. SCOTT: BAHIA GRASS. 113 bed and set out in rows 24 inches wide, with the plants 24 inches apart in the row. Two rows, each about 200 feet long, were planted. The plat was fenced to keep off the stock. The grass made a good growth the first summer, and a good crop of seed was produced the first sea- son. The latter part of September, 191 5, the fence was removed and cattle have pastured on it both winter and summer since that time. During the past four years, the grass has been subjected to heavy pasturing. Notwithstanding this fact, it has continued to grow and has made a complete sod over a space now 10 to 12 feet wide. In addition to this, a large number of individual plants have sprung up adjacent to the planting, indicating that the seed has been scattered by the cattle, birds, or wind. This is, w r e believe, a new method of testing the value of a grass for pasture. A grass to be of value for pasture should, in addition to being nutritious, have good staying qualities. That is, it must stand hard and close pasturing under all conditions. A grass that needs to be nursed and coaxed after it is once established is not desir- able for pasture purposes. The method here employed gives information on two important points, namely, the ability of the grass to spread and make a good sod while being pastured, and its palatability to cattle. The results of this test show that Bahia grass will spread and make a complete sod under pasture conditions. It has also shown that cattle like this grass, as they graze on it at all seasons of the year. Bahia grass seems best adapted to a rather moist soil. This does not necessarily mean a low, poorly drained soil, but rather one that holds moisture well. However, it has been grown on rather dry, sandy soil on the experiment station grounds with fairly satisfactory results. It is not likely to be of any value when planted on dry, sandy ridges. Neither is it likely to be a success when grown on land that is subject to overflow, especially where the water stands for several days. The original seed plat and the plat in the pasture are both on land which is ordinarily considered first class farm soil in Florida. Bahia grass is rather sensitive to cold. A temperature of 34 to 26 degrees will nearly always kill all green growth of this grass. The roots apparently are not' injured by frost or light freezes. When moisture conditions are favorable, growth starts in the spring about the same time as other perennial grasses. No commercial seed of this grass is yet available, but efforts are being made to establish a supply. It seeds freely in Florida when not pastured. When once established it should not be a difficult matter to gather the seed for additional planting. 114 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. AGRONOMIC AFFAIRS. MEMBERSHIP CHANGES. Thanks to the cooperation of certain members, the Society has shown a healthy growth during the past month. The membership reported in the February issue was 504. Since that time, 19 new members have been added and 2 lapsed members have been reinstated, while 1 member has resigned, making a net gain of 20 and a total membership at this date of 524. This is far less than it should be, however, and the further cooperation of the members in bringing the Society to the attention of agronomic workers is urged. The names and addresses of new and of reinstated members, the name of the member who has resigned, and such changes of address as have come to the attention of the officers, follow. New Members. Albertz, H. W., Agronomy Bldg., Univ. of Wis., Madison, Wis. Bastian, Elias, Brookings, S. Dak. Bonnett, O. T., County Agr. Agent, Blue Rapids, Kans. Brown, B. A., Storrs, Conn. Buie, T. S., Experiment, Ga. Carter, C. E., Dept. of Field Crops, Agr. College, Columbia, Mo. ' Corwin, Walling, Dept. of Field Crops, O. S. U., Columbus, Ohio. Homewood, S. L., State College, West Raleigh, N. C. Jensen, Ward C, Clemson College, S. C. Landon, Ira K., 151 i Leavenworth Ave., Manhattan, Kans. Leppan, H. D., Transvaal University College, Pretoria, S. Africa. Letson, Orrin W., Dept. of Farm Crops, Agr. College, Columbia, Mo. Lutz, Dexter N., Dept. of Farm Crops, O. S. U., Columbus, Ohio. May, Ralph W., Judith Basin Substation, Moccasin, Mont. Olinger, R. F., County Agr. Agent, Marion, Kans. Suttle, A. D., College Station, Texas. Tanner, E. L., College Station, Texas. Vass, A. F., University of Wyoming, Laramie, Wyo. Yost, T. F., 830 Fremont St., Manhattan, Kans. Members Reinstated. Hunnicutt, B. H., Lavras, Minas, Brazil. Morgan, J. O., Experiment Station, College Station, Texas. Member Resigned. Wheeler, Clark S. Changes of Address. Cardon, P. V., Agr. Expt. Station, Bozeman, Mont. Fahrnkopf, H. F. T., Farm Bureau, Bloomington, 111. AGRONOMIC AFFAIRS. Grantham A. E., Ya.-Car. Chemical Co., Richmond, Va. Kraft, J. H., College Station, Texas. Le Worthy, G. E., Harrington, Del. Petry, E. J., 1218 Catherine St., Ann Arbor, Mich. Whitcomb, W. O., Experiment Station, Bozeman, Mont. White, C. L., Aurora, Mo. NOTES AND NEWS. Jay A. Bonsteel, in charge of special soil studies in the Federal Bureau of Soils for the past several years, has resigned to engage in farming in New York. John Buchanan, who has been in charge of cooperative experiments in agronomy at the Iowa station and secretary of the Iowa Crop Im- provement Association since 191 2, has resigned to become county agricultural agent in Story County, Iowa. O. D. Center, formerly director of extension in Oregon, is now county agent in McLean Co., 111., with headquarters at Bloomington. C. C. Cunningham, in charge of cooperative experiments in agron- omy at the Kansas station, has resigned effective March 1 to engage in farming. W. H. Dalrymple, vice-director of the Louisiana station and head of the department of veterinary science in the Louisiana Agricultural College, has been made dean of the college of agriculture and director of the experiment station. George E. Eggington, seed analyst at the Colorado station, has ac- cepted a position with the Minneapolis office of Albert Dickinson Company, wholesale seed dealers. F. E. Fuller, extension agronomist in the Montana college, has re- # signed to accept the position of agricultural adviser to the farm bu- reau of Marshall and Putnam counties, Illinois. A. C. Hartenbower, superintendent of farmers' institutes and extension schools in Oklahoma, has resigned to engage in farming. Harry Hayward, dean of the Delaware college of agriculture and director of the experiment station, has resigned to join the staff of N. W. Aver & Son, a Philadelphia advertising agency. He has been succeeded by C. A. McCue, formerly horticulturist of the Delaware college and station. Il6 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. David F. Houston, secretary of agriculture since March, 1913, on February 2 succeeded Carter Glass as secretary of the treasury. He was succeeded as secretary of agriculture by Edwin T. Meredith, of Des Moines, Iowa, owner of Successful Farming, a director of the Chicago Federal Reserve Bank, and president of the Associated Advertising Clubs of the World. George Livingston, acting chief of the Federal Bureau of Markets since the resignation of Charles J. Brand last summer, has been made chief of the Bureau. H. C. Ramsower, formerly professor of agricultural engineering in Ohio State University, is now director of extension in Ohio. Albert Osenbrug, recently in charge of dry-land agriculture experi- ments at the Colby, Kans., substation, has been made superintendent of the Judith Basin Substation, Moccasin, Mont. C. C. Ruth, formerly of the Portland, Oreg., office of grain stand- ardization, Federal Bureau of Markets, is now assistant agronomist at the Oregon station, succeeding F. S. Wilkins, now of the Iowa station. Herschel Scott, formerly assistant in agronomy at the Kansas col- lege, resigned March 1 to engage in commercial work in California. M. C. Sewell, assistant professor of soils in the Kansas college, is on leave for graduate study in the University of Chicago. W. O. Whitcomb, recently in charge of the Minneapolis office of the seed-reporting service, Bureau of Markets, is now in charge of the seed laboratory of the Montana experiment station. JOURNAL OF THE American Society of Agronomy Vol. 12. April. 1920. No. 4. THE STATUS OF LIME IN SOIL IMPROVEMENT. 1 Elmer O. Fippin. The use of lime on the soil is shown by practical experience to be needed by such large areas and the scientific questions involved with its correct and economical use are revealed to be of such complicated and far-reaching character that the further investigation of this sub- ject is a matter of major importance in crop production and soil improvement. The underlying scientific reasons for this need for lime and the functions performed by it in the soil and in the plant are still matters of wide differences of opinion among investigators. This discussion ranges over the questions : Is there free acidity in the soil? What is the relation of free acid to the lime or other base-absorption coeffi- cient of the soil ? What tests, if any, constitute an adequate measure of the need for lime by a particular soil for the growth of so-called acid-sensitive crops? Is free acidity in itself the limiting factor or is it correlated with some other condition which is responsible for the character of plant growth, such as the presence of aluminum nitrate ? The opinion is quite general among investigators that the need for lime is associated either in a direct or an indirect way with an acid condition of the soil as measured by the absorption of a base, upon which principle rests most of the methods of measuring the need for lime. THE USE OF LIME MATERIALS IN THE SOIL. Concerning the range of tolerance by different crops of an acid condition of the soil, very little is definitely known, but a wide varia- 1 Presented at the twelfth annual meeting of the American Society of Agron- omy, Chicago, 111., November 10, 1919. 117 I 1 8 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. tion is indicated, for example, by the distinctions between alfalfa and blueberries or red sorrel and red clover. The sorrel plant (12, 14) 2 illustrates a further fact, namely, that some plants have a wide range of tolerance of both an acid and a lime-rich condition of soil, while other plants may have a narrow range of tolerance. The range of tolerance of microscopic plants such as those con- cerned with the transformation of nitrogen and those concerned with the production of a diseased condition of plants, such as the potato scab (5) and the club root of cabbage, is equally important from the viewpoint of farm practice. The lack of accurate information con- cerning the tolerance of the lower plant forms is equally as great as it is concerning the higher plants. Certain it is that plants cannot be divided sharply into two classes, one of which will thrive only on an alkaline soil while the other will thrive only on an acid soil. We believe that every graduation of tolerance is exhibited among dif- ferent plants. Herein arises another important point. Too often if has been assumed that for plants that thrive on a soil near the neutral point too much lime carbonate could not be present in the soil. The question might very properly be asked whether the alkaline or calcium-magnesium tolerance of plants may not be quite as important to determine as their tolerance of the opposite or so- called acid condition. The investigations of Fred 3 of Wisconsin, the studies of chlorosis (4) or inability to absorb iron in certain lime-rich soils in Florida, and field observations by the writer and others point to the impor- tance of this subject. EFFECTS OF LIME ON SOIL. Coming now to the use of lime on the soil, including both the caustic and the carbonate forms, two classes of problems arise, namely, (a) What are the relative effects of equal amounts of the oxides of calcium and magnesium on the chemical, physical, and bio- logical properties of the soil? and (b) What are the relative practical aspects of the use of these different materials? Several investigations are in progress on the effect of liming mate- rials on the chemical nature of the soil and the soil solution, under laboratory, plat, and field conditions. These have not reached a con- clusive stage, as is illustrated by the data on the relation of lime to the availability of phosphorus and to a less extent of potassium (1, 3). Equally undetermined is the ultimate relation of lime to the 2 Reference is to " Literature cited," p. 123. 3 Personal communication to the writer. fippin: lime in soil improvement. 119 store of nitrogen in the soil, especially when its use is combined with the growth of a legume. Much misinformation has been given out on the effect of different forms of lime on the disappearance of organic matter from the soil. Here, distinction has not been made between purely chemical effects of the lime compound on the organic matter and the biological effects resulting from the stimulation of the growth of microorganisms in the soil by lime materials. The growth of such organisms is inevi- tably at the expense of the organic matter in the soil. To what ex- tent is this effect necessary and legitimate and to what extent may it be undesirable? Is it a similar effect for both carbonate and caustic forms of lime? The statement, common in the older agricultural literature (8), that caustic lime applied to the soil in even reasonable amounts is especially destructive of organic matter, has usually been put in a form to indicate that this destruction is a purely chemical process, such as occurs when spontaneous combustion of inflammable material results from the contact of a large amount of water with a consid- erable quantity of quick lime. This idea of the destruction of organic matter has extended to hydrated lime, because it also has caustic properties, but it has no capacity for chemical union with water in- volving the liberation of heat. Further, the slacking of lime in the soil, say as granular quick lime, cannot result in the rise of tempera- ture necessary to a destructive chemical change. Unquestionably there is chemical union of the lime with constituents of the organic matter. That this union is truly destructive of the organic substance, as would be indicated by the liberation, even in strongly alkaline solu- tions, of carbon dioxide, has not been demonstrated and the phe- nomena are not in accord with the known principles of organic chemistry. The chemical and biological relations of this problem must be kept clearly separated. If organic matter decomposes more rapidly where caustic forms of lime have been used than where carbonate forms have been applied, as is frequently claimed, it raises the question whether, as a result of these more active chemical and biological reac- tions, the use of caustic forms of lime in suitable amounts may be better than the use of carbonate forms. Who can say what are the relative effects of caustic and carbonate forms of lime on the granu- lation and on the porosity and related properties of different soils? Are these effects the same or do they vary with different kinds of soil? Available data indicate that in any direct way caustic 120 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. forms of lime have the largest granulating effect (2) on clay soils, while carbonate forms are either nearly inactive or produce positively an unfavorable physical change. Do the available data on this point furnish an adequate guide? Closely connected here is the question of how long caustic lime re- mains in the hydrated form in the soil, and into what new combina- tions either the caustic or carbonate form enters in the soil. Maclntire (11) and Mooers of the Tennessee station have done much work showing, first, that caustic forms of lime are not chemically destruc- tive of organic matter; second, that recarbonation proceeds very rapidly and is normally completed in a few days at the outside ; and third, that magnesium and to a less degree high calcium limes rap- idly enter in silicate combinations and that these new combinations markedly affect the solubility and movement of those constituents in the soil. Especially does the magnesium seem to increase the move- ment of sulfur. Conner 4 of Indiana has data showing that calcium in silicate combination, as in basic slag, may be nearly as effective in performing the functions of lime in the soil as when applied in caustic or carbonate form. FORMS AND FINENESS. This matter of the value of lime in certain types of silicate and similar combinations is particularly important because it is related to the matter of fineness of lime materials applied to the soil. If lime in these silicate forms of combination is just as effective in affecting the yield of crops as if it were in carbonate form, it is then quite as permissible to apply those forms of lime that enter most actively into these new combinations, namely, burnt lime and finely pulverized carbonate, as to use the more inactive coarse carbonate. There may even be an advantage from the formation of these sili- cates because, first, they suggest the precipitation of colloidal sili- cates ; second, they maintain a more mild alkalinity ; and third, they aid in conserving the lime materials in the soil without interfering with their usefulness. Growing out of this same question of form and fineness is the ques- tion of the movement of lime thru the soil and the possibility of loss from leaching. The lysimeter data (10) collected at Cornell Univer- sity during five years do not show any increase in loss of calcium and magnesium when lime was applied. These short-period data are vitiated by the existence of a large amount of limestone in the deep subsoil, which would tend to mask any movement from the surface 4 Personal communication to the writer. FIPPIN I LIME IN SOIL IMPROVEMENT. 121 soil. Maclntire has found the leaching of lime thru a deep section of soil to be essentially independent of the rate of application in any ordinary period of a few years. Let it be noted here that we are not concerned with what may happen in a thousand years, but with what is the practical loss from the vertical soil section that will occur in three, five, or ten years, which is as long a period as the applica- tion of lime is intended to cover. The analysis of the soil of one of the fields at Rothamsted (6) shows the presence of as much as 3.3 percent of carbonate of lime in the surface 9 inches and none in the second 9 inches. This carbo- nate of lime seems to be the result of application of chalk so long ago that the record is lost. Its persistence in the soil and the lack of movement into the subsoil indicate how slow is the movement of lime materials. We come now to the question of suitable fineness of limestone. We have largely disposed of the question of extreme fineness. The next question here is, how large may particles of lime carbonate be and still perform the full functions of such material in the period for which it is applied, namely, from three to six years or for an average rotation? First of all, let us be reminded that certain processes in the soil are inhibited, as compared with their operation in a free liquid. This is especially true of diffusion. Lime carbonate is solu- ble in an acid solution, and will continue to dissolve as long as the acid is present in contact with the material. In the soil, the question arises, thru how wide an area of soil does diffusion operate when this soiLis in the optimum moisture condition? From most of the investigations available it seems to be very slow, and to reach a very short distance from the point of solution. If there is more lime in a particle than is required to neutralize the acid solution or satisfy the lime-absorption coefficient of the soil within the active range of the surface of the lime particle, then at the end of this reaction the re- mainder of the particle of lime will be essentially sealed in a shell of the alkaline soil where it may remain for a long period except as it is disturbed by mechanical means. How coarse may a particle be before this condition occurs in the average acid soil? Is the maxi- mum size of such a particle as large as one fourth or one tenth of an inch in diameter, or is it down around one fiftieth to one eightieth of an inch in diameter? The practical data on this point are very meager. Observations on calcareous glacial soils reveal particles of carbonate of lime in soils the greater part of which are distinctly acid to litmus and which respond with larger crop growth where lime is 122 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. applied. Experimental field data (7, 9, 13, 15) are available for so short a period or have been secured under such conditions of soil, crop succession, and rate of application as to make them of question- able value as a guide in this practical matter. Certain it is that such studies are not adequately conducted unless four conditions are met : 1. The soil must be distinctly in need of lime thruout a vertical section at least 4 feet deep from the surface. 2. The limestone must be sorted into' rather narrow textual divi- sions and used in oxide-equivalent amounts. 3. The rates of application should range from a very small quan- tity, such as 200 or 300 pounds, up to as large a quantity as several tons. 4. The crop should be one sensitive to an acid condition and one not able to succeed in that soil without lime. A fifth condition may be added, namely, insurance that there is an adequate supply of nutrients such as phosphorus. The question of suitable fineness cannot be regarded as settled in any sense, nor is it sufficient to> advise the use of a large quantity of coarsely ground material on the chance that there may be enough fine material to sup- ply the needs of the soil for good plant growth. This runs into eco- nomic questions. A further point involved is the extent to which the time element may compensate for lack of fineness. FIELD EXPERIMENTS AT THE PRESENT TIME. Almost none of the field experiments involving the study of lime materials is designed in a way adequately to investigate any one or more of the important scientific and practical problems involved in the use of liming materials on the soil. Certainly no adequate data of that sort have accumulated. This is not meant to cast any undue reflection on such carefully maintained or long continued work as that at the Ohio, Pennsylvania, and other stations. In the planning of those experiments the natural human limitation attaching to the investigations in a new field has been involved. RELATION OF FORM OF LIME TO THE TYPE OF SOIL. The questions of caustic vs. carbonate lime, fine vs. coarse lime, calcium vs. magnesium, and the amount of lime necessary for par- ticular crops have not been settled by scientific investigations and for the guidance of practice must rest largely on the empirical results of field trials. Such empirical or practical field data as well as real experimental work should include results obtained on a number of FIPPIN I LIME IN SOIL IMPROVEMENT. 123 types of soil. It is not safe to draw conclusions from results on a single type of soil. Finally, the practical aspects of these questions are embodied in laws covering the sale of liming materials, which are now widely divergent in procedure and requirements, and reflect the unsettled state of the general knowledge on the subject. Certainly, the essen- tial elements of a lime-inspection law must be very much the same in the different States. If it is oxides of calcium and magnesium with which the farmer is concerned, then the first step would seem to be to report all forms of liming materials, both carbonate and caustic, on the basis of their oxide content. If fineness is a factor in value, then the fineness of carbonate materials should be reported for a standard series of screens ranging in size from the coarsest material that has appreciable value down to as fine materials as seems to be of importance under field conditions. The United States Bu- reau of Standards has recently promulgated a new system of speci- fications for testing screens, and all provision for screen analysis should be in harmony with these specifications and uniform among the different States. I hope that the information, energy, and facilities of all the workers in the field of soil and crop improvement may be pooled in some kind of a broad conference to study all these questions, and as rapidly as possible to standardize our information and practices with reference to the use of lime in soil. Literature Cited. 1. Briggs, L. J., and Breazeale, J. F. Availability of potash in certain ortho- clase-bearing soils as affected by lime and gypsum. In Jour. Agr. Re- search, v. 8, no. 1, p. 21-28. 1917. 2. Fippin, Elmer O. Some causes of soil granulation. In Proc. Amer. Soc. Agron., v. 2 (1910), p. 106-121. 1911. 3. Gaither, E. W. Effect of lime upon the solubility of soil constituents. In Jour. Indus, and Engin. Chem., v. 2, no. 7, p. 315, 316. 1910. 4. Gile, P. L. Relation of calcareous soils to pineapple chlorosis. Porto Rico Agr. Expt. Sta. Bui. 11. 191 1. 5. Gillespie, Louis J., and Hurst, Lewis A. Hydrogen-ion concentration — soil type — common potato scab. In Soil Sci., v. 6, no. 3, p. 219-236. 1918. 6. Hall, A. D. Book of the Rothamsted Experiments, 2d ed., p. 140. E. P. Dutton & Co. 1917. 7. Hartwell, B. L., and Damon, S. C. A field comparison of hydrated lime with limestone of different degrees of fineness. R. I. Agr. Expt. Sta. Bui. 180. 1919. 8. Hopkins, C. G. Soil Fertility and Permanent Agriculture, p. 162. Ginn & Co. 1910. 124 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. 9. , Garrett, F. W., Whitchurch, J. E., and Fahrnkopf, H. F. T. Illinois crop yields from soil experiment fields. 111. Agr. Expt. Sta. Bui. 219. 1919. 10. Lyon, T. L., and Bizzell, J. A. Lysimeter experiments. N. Y. (Cornell) Agr. Expt. Sta. Memoir 12. 1918. 11. MacIntire, W. H. The carbonation of burnt lime in soils. In Soil Sci., v. 7, no. 5, p. 423-436. 1919- 12. Pipal, F. J. Red sorrel (Rume.v acetosella) and its control. Ind. Agr. Expt. Sta. Bui. 197. 1916. 13. Stewart, R., and Wyatt, F. A. The comparative value of various forms of limestone. In Soil Sci., v. 7, no. 4, p. 273-278. 1919. 14. White, J. W. Concerning the growth and composition of clover and sorrel (Rumex acetosella) as influenced by varied amounts of limestone. In Ann. Rept. Pa. State College for 1913-14, part 2, p. 46-64. 1915. 15. The maintenance of soil fertility; a quarter century's work with manure and fertilizers. Ohio Agr. Expt. Sta. Bui. 336. 1919. THE INHERITANCE OF RESISTANCE TO BUNT OR STINKING SMUT OF WHEAT. 1 E. F. Gaines. The material presented in this paper concerns the bunt resistance of three wheats, Turkey (Washington No. 326), Hybrid 128 (Wash- ington No. 592), and Florence (Washington No. 634), and the resist- ance of the F 2 and F 3 as well as selections in the F 4 generation of two crosses, Turkey X Hybrid 128 and Turkey X Florence. The wheats under investigation at the Washington Agricultural Experiment Station show great variation in susceptibility to bunt (Tilletia tritici). The comparative resistance of thirteen different varieties under conditions of maximum infection is described in an earlier article. 2 DESCRIPTION OF PARENTS. According to the previous report, Turkey (T. sativum vulgare) produced 1.81 percent of bunt, whereas Hybrid 128 (T. sativum com- p actum) with similar conditions of infection produced 92.15 percent. These two varieties are of considerable importance in Washington, about 2,000,000 bushels of Hybrid 128 and 1,000,000 bushels of Tur- 1 Contribution from the Washington Agricultural Experiment Station, Pull- man, Wash. Received for publication February 9, 1920. 2 Gaines, E. F. Comparative smut resistance of Washington wheats. In Jour. Amer. Soc. Agron., 10, no. 3, p. 218-222. 1918. GAIXES: RESISTANCE TO WHEAT BUNT. 125- key being grown in 1918, which was approximately 40 percent of the winter wheat produced in the State that year. Notwithstanding its great susceptibility to bunt, Hybrid 128 is so prolific, so winter hardy, has such stiff straw, and holds the grain so well after maturity that it is grown in larger quantities than any other winter wheat in Washington. The grain weighs well and grades commercially as white club. If it were not for its suscepti- bility to bunt, it would be grown by many farmers who now sow other varieties. Turkey wheat is drouth resistant, winter hardy, of high milling quality, bunt resistant, and grades as hard red winter. It is produced in considerable quantities in the semiarid sections of Washington in years that are favorable for winter wheat. The straw is not so stiff as that of Hybrid 128, the grain shatters more easily, and the beards are objectionable. Florence is a spring wheat of Australian origin, grading as com- mon wmite, which has also proved to be highly resistant to bunt. It is not grown commercially in Washington. RESISTANCE OF PARENTS AND HYBRIDS COMPARED. In the summer of 191 5, several hybrids were made between Turkey and Hybrid 128, and between Turkey and Florence. In 1918, a complete F 3 of each of these crosses was analyzed for bunt resistance. The parent varieties and F 2 sibs were grown in the same plat with the F 3 rows under the same conditions and in the same year, for more direct comparison. Turkey and Florence both showed a high degree of resistance, but Hybrid 128, altho very susceptible, did not smut so badly as indicated by the 4-year average of the same variety. The F 2 of Turkey X Hybrid 128 showed an infection intermediate be- tween the two parents. The F 2 of Turkey X Florence produced a much higher degree of bunt than either parent. In this cross the lack of winter hardiness on the part of Florence made it difficult to get large numbers. Table 1 shows the actual numbers as they were counted in the field and is the basis of the percentages given in Table 2. The data presented in Table 1 are all taken from field records. The material was grown on land that had been in peas the year before. Approximately a half gram of bunt spores was added to each packet of seed before sowing, in order to obtain maximum in- fection. The sowing was all done the same week and at a uniform depth. The soil was uniform and in good condition. The rows 126 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. were 18 inches apart and the seeds were spaced 6 inches apart in the row. During the spring and summer, enough cultivation was given to keep down weeds. At harvest time the plants of each row were pulled and divided into three classes, bunt-free plants, partly bunted plants, and entirely bunted plants. The number in each class was then recorded. The partly bunted plants were further divided into heads of wheat and heads of bunt, and the number of each recorded. In case a head was partly bunted it was placed with the pile it resem- bled most. Table i. — Number of plants bunt free, partly bunted, and entirely bunted in 1918, with the number of uninfected and bunted heads produced by the partly bunted plants. Variety or hybrid. Number of plants. Number of heads on partly bunted plants. Bunt free. Partly bunted. All bunted. Total. Not bunted. Bunted. Turkey 485 17 502 191 41 Hybrid 128 43 46 59 148 228 404 Florence 81 5 86 37 II Turkey X Hybrid 128, F > 45 93 53 191 678 613 Turkey X Florence, F2 143 3i 5 179 33i 188 Turkey X Hybrid 128, F 3 3.678 4,094 3,814 11,586 27,194 26,223 Turkey X Florence, F3 1,603 257 199 2,059 2,282 1,902 To get the percentage of infection, the number of partly bunted plants and those entirely bunted were added and the sum divided by the total number of plants in the row. In figuring the percentage of bunted heads on the infected plants, it was assumed that the plants which were entirely bunted had the same average number of heads as the partly bunted plants, the num- ber of the latter being determined by actual count. To get the total percentage of bunt produced, the percentage of totally bunted plants was added to the percentage of bunt produced by the partly bunted plants when figured in terms of the whole row. The number of rows indicate the number of replications of the parent varieties and F 2 sibs, and show the size of each F 3 family. It is understood that each F 3 row is the product of an F 2 plant, grown in 1 91 7. All the F 3 rows of each cross are the progeny of a single F 1 plant produced in 1916. The F 2 rows grown in 1918 were from 2-year-old seed of Fj. plants which were full sisters of the F x an- cestors of the two F 3 families. The data are presented in Table 2. GAINES: RESISTANCE TO WHEAT BUNT. 127 Table 2 shows the high degree of resistance of the Turkey and Florence varieties. I Ivbrid 128 is less susceptible to bunt than usual, but all varieties showed an abnormally low percentage of the disease in 191 8. Not only do Turkey and Florence show a much lower in- fection than Hybrid 128, but the infected plants themselves produce much less bunt than the infected plants of Hybrid 128. This is sig- nificant in itself, as it suggests the possibility that there are two kinds of resistance, one which prevents infection and one which retards the development of the fungus after infection. Table 2. — Percentage of infected plants, percentage of bunt on infected plants, and total percentage of loss from bunt in 1918 of the three parents and F 2 and F s generations of Turkey X Hybrid 128 and Turkey X Florence. Bunted Average Variety or hybrid. Infected heads on Total bunt Number number of plants. infected produced. of rows. heads per plants. plant. Percent. Percent. Percent. Turkey 3-4 17.7 0.6 6 13.6 Hybrid 128 70.9 63.9 59-8 3 13-7 5-8 22.9 1-3 3 9.6 Turkey X Hybrid 128, F 2 76.4 47 9 51-0 3 13-9 Turkey X Florence, F2 20.1 35-5 8.9 2 16.7 Turkev X Hybrid 128, F 3 68.3 49.1 50.2 194 13-0 Turkev X Florence, F3 22.1 45-5 15-4 168 16.3 The F 2 of Turkey X Hybrid 128 shows a higher infection than the susceptible parent, but a much lower percentage of bunt is produced on the infected plants. The latter condition is to be expected as an intermediate condition, but it is difficult to explain the former except as a parallelism of increased hybrid vigor, which might cause more of the infected plants to survive the winter. The average of the F 3 bears out this supposition, for there is a slight reduction in the per- centage of infected plants and a very slight increase in the percentage of bunt on the infected plants. The F 2 of Turkey X Florence has a much higher degree of infec- tion than either parent and also a higher percentage of bunt on the infected plants. The average of all F 3 rows shows a slight increase in the percentage of infected plants over the F 2 and a decided in- crease in the percentage of bunt on the infected plants. If resistance is caused by a series of chemical or physiological factors, then two facts are self evident. Most of the factors for resistance are recess- ive, and Turkey and Florence each have different and distinct factors for resistance. 128 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. AVERAGE RESISTANCE, IN F y , OF DIFFERENT MORPHOLOGICAL SEGREGATES. Table 3. — Distribution of bunt infection among the different- morphological types segregating in the F z of Hybrid 128 X Turkey. Number Plants Plants Plants part bunt. Total Spike characteristics. bunt Bunt heads. number of rows. free. 3. 1 1 bunt Number. Wheat heads. of plants. Long bearded 7 199 187 131 835 1,096 517 Long mixed 24 549 505 508 3.399 2,714 1,562 Long beardless 6 105 165 63 374 418 333 Mixed bearded 13 355 214 346 2,626 2,041 915 Mixed mixed 52 863 1,065 1,267 8,334 7,703 3,195 Mixed beardless 24 39i 382 472 3.027 3.297 1-245 Club bearded 14 3ii 198 307 2,090 2,013 816 Club mixed 33 587 610 676 4.479 4,346 1.873 21 318 488 324 2,030 2.595 1,130 Total 194 3,678 3.814 4.094 27,194 26,223 11.586 In Table 3, the F 3 of Turkey X Hybrid 128 is arranged according to the characteristics of head length and floral-glume appendage. The distribution of the nine types is quite abnormal for a dihybrid. There are far too many rows homozygous for the club character and also for the beardless character. This is a common variation in certain crosses of this strain of Turkey, but it is outside the province of this paper to discuss this point. Table 4. — Distribution of bunt infection among the different morphological types in the F :i of Hybrid 128 X Turkey, figured in percentages. Spike characteristics. Number of rows. Plants bunt free. Plants all bunt. Plants part bunt. Bunt heads on partly bunted plants. Total bunt. Average plants per row Percent. Percent. Percent. Percent. Percent. Long bearded 7 38.49 30.17 25-34 56.75 50.55 73-9 Long mixed 24 35-15 32.33 32.52 44.40 46.76 65.1 Long beardless 6 31-53 49-55 18.90 52.78 59-52 55-5 Mixed bearded 13 38.80 23.38 37-80 43-73 39-90 70.4 Mixed mixed 52 27.00 33-30 39-65 48.03 52.34 61.4 Mixed beardless 24 31.40 30.68 37-90 52.13 50.43 5i-9 Club bearded 14 38.10 24-30 37-62 49.06 43-26 58.3 Club mixed 33 31-34 32.57 37.00 49.24 50.79 56.8 Club beardless 21 28.14 43.18 28.67 56.11 59-27 53-8 Total 194 31-75 32.92 35-34 49.01 50.27 60.8 Table 4 presents the data of Table 3 in terms of percentages for comparison. There is a wide difference in the total percentage of bunt produced by the different types. The club beardless type is highest (59.27 percent) or almost identical in total bunt with the sus- GAINES : RESISTANCE TO WHEAT BUNT. I2 9 ceptible parent of the same type. The rows that had mixed long and club bearded heads produced the smallest amount of bunt (39.90 percent). The correlation of bunt resistance and morphological characters breaks down, however, when all bearded types are added together and compared with the sum of the beardless types and a similar comparison made between the sum of all long and all pure club types. The slightly higher resistance of the bearded types is well within the limits of error to be expected in an investigation of this kind. Table 5. — Distribution of bunt infection among the bearded, mixed, and beard- less F 3 rows of Turkey X Florence, as shown by actual count of bunt free, all bunted, and partly bunted plants. Spike characteristics. Number ot rows. Plants bunt free. Plants all bunt. Plants part bu XT . Wheat Number. heads> nt. Bunt heads. Total number of plants. Bearded Mixed Beardless 38 78 52 387 795 421 74 86 39 90 112 55 823 1,017 442 585 892 425 551 993 515 Total 168 1.603 199 257 2,282 1,902 2,059 The distribution of types in the F 3 of Turkey X Florence shows the same prepotency of the beardless type. The total number of plants, however, suggests greater winter hardiness of the bearded types. Table 6. — Distribution of bunt infection among bearded, mixed, and beardless Ft rows of Turkey X Florence, in percentages. Spike characteristics. Number of rows. Plants bunt free. Plants all bunt. Plants part bunt. Bunt heads on partly bunted plants. Total bunt. Number of plants per row. Percent. Percent. Percent. Percent. Percent. 38 70.2 13-4 16.3 41-5 20.2 14-5 Mixed 78 80.1 8-7 II-3 46.7 14.0 12.7 Beardless 52 81.7 7.6 10.7 49.0 12.8 9.9 Total 168 77-9 9-7 12.5 45-5 15.4 12.6 According to Table 6, there is an inverse correlation between the bearded characteristic and bunt resistance. In Table 2, Turkey was shown to be twice as resistant as Florence, yet in the F 3 the bearded rows produced nearly twice as much bunt as the beardless rows. In Table 4, the long bearded type produced almost 4 percent more bunt 130 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. than the long beardless type, which is a variation in the same direc- tion. With the club types, however, this variation is reversed, and the mixed rows for head length as well as the pure club rows show a smaller percentage of bunt in the bearded than in the beardless segregates. Incidentally, very resistant rows, very susceptible rows, and numer- ous gradations between these extremes were found in each morpho- logical type. Probably the large variation in the types shown in Ta- bles 4 and 6 are accidental variations due to the numerous variants under consideration. SEGREGATION IN THE F 3 ON THE BASIS OF BUNT RESISTANCE. Table 7. — Distribution of bunt in the different F 3 rows of two crosses which show the genetic segregation of the F 2 plants, the first being the result of crossing two resistant varieties, while the second is the result of crossing a resistant variety on a susceptible variety. Total bunt. Rows of F3, Turkey X Florence. Rows of F3, Turkey X Hybrid 128. Total bunt. Rows of F3, Turkey X Florence. Rows of F3, Turkey X Hybrid 128. Percent Percent. None 72 20 to 29.9. . . . 20 9 0.1 to 4.9. . . . 14 4 30 to 39-9 12 23 5.0 to 9.9 15 10 40 to 59.9- • • • 12 58 10. to 14.9. . . II 5 60 to 79.9. . . . 4 64 15.0 to 19.9. . . 9 8 80 to 99.9. . . . 2 13 Table 7 shows that bunt resistance is definitely heritable in both crosses, tho the distribution is very different. The 72 immune rows of Turkey X Florence may vary greatly in the amount of resist- ance they contain, but, since they were all immune, the only way to isolate the differences would be to hybridize each with a susceptible variety and analyze the F 3 . This will take time and patience. It is evident, however, that the resistance of the two parents are cumula- tive in effect, for the parents produced an average of 4.6 percent of infected plants under the same conditions that the 72 immune rows were grown. The 50 rows that produced more than 20 percent of bunt indicate that the resistance of at least one parent is composed of multiple factors, but since between 3 and 4 percent show no re- sistance (that is, they produced as much bunt as any of the suscepti- ble varieties) it is not likely that there are more than four or five factors affecting resistance in this cross. The F 3 of Turkey X Hybrid 128 produced a continuous series from one row as resistant as Turkey to 77 which were more sus- ceptible than Hybrid 128. The very small number of highly-resistant GAINES : RESISTANCE TO WHEAT BUNT. I 3 I rows indicates a dominance of the susceptibility of Hybrid 128. There is evidently no single or latent heritable factor for resistance in Hybrid 128, for not one out of 194 F 3 rows showed more resist- ance than Turkey. On the other hand, there must be at least three independent factors for resistance in Turkey. Otherwise, there would have been a larger number of rows as resistant as Turkey. The increased susceptibility of so many rows is difficult to explain, for it is far beyond that of Hybrid 128 grown under similar conditions. Table 8. — Inheritance of bunt resistance in the F 4 (1919) generation of selections made from the most resistant F :i rows. TURKEY X HYBRID 128. Row No. Number of rows in F4. Average num- ber of plants per row in the F 4 . Bunt produced in the F4, average of all rows. Bunt produced in the F 3 parent row, 1918. Least bunty row in the F4. Buntiest row in the F<. Percent. Percent. Percent. Percent. I 27 63 13-2 7.0 4-3 26.7 2 26 71 12.7 1.9 5-6 35-0 22 56 18.7 4.8 6-5 39-6 4 17 46 34-8 14.6 15-5 54-7 5 14 42 33-2 9.8 11.6 59-5 6 14 66 13-6 6.6 5-6 25.4 7 12 70 25-3 7-1 12.7 40.3 8 8 69 26.6 13-4 16.6 41.7 9 2 50 32.0 6.4 2 5 -9 38.1 TURKEY X FLORENCE. I 12 59 .02 •3 2 11 54 .02 .2 11 44 •97 3-2 4 7 27 1.97 5-5 Selections were made from nine of the most resistant rows of the compaction type for testing Turkey X Hybrid 128 in the F 4 . The results shown in Table 8 indicate that none of the 9 F 3 rows were homozygous for all of the resistant factors. Neither were the 112 F. row r s as susceptible as Hybrid 128, which produced 98.1 percent of bunt in 1919. Turkey produced 5.9 percent under the same condi- tions. In fact, all the principal varieties produced a very high per- centage of bunt in 1919, which shows that the season was unusually favorable for bunt. With this explanation it would seem that the 1919 record of Table 8 is not so high from an inheritance standpoint as the figures would indicate. The first and sixth selections were probably homozygous for two or more of the units of resistance, for all of the F 4 rows showed a high degree of resistance. 132 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. The 41 selections from the 4 rows of Turkey X Florence were all remarkably resistant in 191 9, 25 being entirely immune. The most susceptible F 4 row of the 41 was more resistant than either parent. Turkey produced 5.9 percent and Florence 8.1 percent of bunt in 1919, whereas the most susceptible F 4 row produced 5.5 percent. DISCUSSION AND INTERPRETATION OF RESULTS. No attempt has been made to find out what resistance is, nor have the conditions of moisture and temperature affecting infection been discussed. The material here presented is on a comparative basis only. The conditions of infection and the time and method of sow- ing were as similar as possible under field conditions. The wide differences in the amount of bunt produced in the F 3 under these conditions in comparison with the constancy of the per- formance of the parent varieties seem to warrant the following conclusions : 1. Bunt resistance in wheat is not a simple Mendelian unit char- acter. • 2. Resistance, if Mendelian, is composed of multiple factors, for a continuous series ranging from complete immunity to complete sus- ceptibility has been obtained. 3. Different wheat varieties possess different kinds of resistance. 4. Linkage between resistance and morphological characteristics is not sufficient to prevent the selection of a resistant strain of any mor- phological type desired. WALDRON : FIRST GENERATION ALFALFA CROSSES. 133 FIRST GENERATION CROSSES BETWEEN TWO ALFALFA SPECIES. 1 L. R. Waldron. This paper deals with the amount of growth made by certain hy- brids between the two species of alfalfa, Medicago sativa and M. f ale at a, in comparison with the parent forms, and also with the amount of winter injury and winterkilling sustained by these hybrids. The experiment as planned also concerned itself with the amount of cross-fertilization occurring among normally pollinated alfalfa plants, as reported in a previous paper (4). 2 Experimental Work. The F 1 plants were started in flats in the greenhouse in January, 1918, the seed having previously been treated with sulfuric acid. The plants were transplanted once and very early in the spring were set out of doors in flats. They were protected from severe freezing by cloth. The seedling plants were transplanted to the field during the first week of May. The soil of the field was a rich, black loam, well drained and in excellent tilth. The plants were set with a hand planter 30 inches apart each way. The progeny of each pistillate parent plant gener- ally comprised 84 or 105 numbers, when the stand was complete. The seedlings planted from the pistillate sativa parent and from the pistillate falcata parent were 2,316 and 2,034, respectively. A very good stand resulted, tho some blank hills were later in evidence. The Fj hybrids from typical plants of the two species of Medicago, such as the parent plants used in this experiment, can be told without hesitancy at time of blooming. The strongly variegated flower color has been described by Westgate (5). The hybrids showed much variation in flower color, but generally only a casual examination was necessary to reveal the characteristic flower color of the hybrid. Notes were taken the first season, 1918, upon each plant as to its 1 Contribution from the North Dakota Agricultural Experiment Station, Agricultural College, N. Dak. Approved by the Director. Presented at the twelfth annual meeting of the American Society of Agronomy, Chicago, 111., November 10, 1919. 2 Reference is to " Literature cited," p. 143. 134 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. hybridity and as to the time of the first appearance of bloom. In the spring of 1 91 9, notes were taken on the amount of winterkilling by noting the number of plants dead and alive. An attempt was also made to judge the amount of winter injury suffered by the plants which persisted. The living plants were classified from 1 to 10, grade 1 being given to the weaker plants and grade 10 to the larger and most vigorous plants. The values thus secured will be consid- ered later. WEIGHTS OF PLANTS. It was not thought possible or desirable to take weights on all of the plants and so a selection had to be made of the plants to be weighed. One had to consider the possible differential effect of the winter upon the hybrid plants and upon their respective sativa and falcata parental forms. 3 As will appear later, there seemed to be a difference of this sort, but it is believed it was well guarded against by the method used in selecting the plants. No plants were selected for weighing which had been given a class value less than 5. All of the inferior plants were thus automatically excluded. Even with these plants excluded, there would be a chance for the sativa plants scaling above 5 to be depressed in value because of the winter influ- ence. In fact, the modal class of the hybrid plants classified as 5 or above lay at 8, with a mean of 7.7 ; the modal class of the sativa plants lay at 7, with a mean value of 6.9. It is obvious that, if the hybrid plants produced more growth than the sativa plants, this fact would naturally be recorded in the data taken as to the amount of winter injury, which we have been discuss- ing. The amount of winter damage was so slight, as a matter of fact, that the vitiation of the experiment from this source was probably negligible. In selecting plants upon which weight data were to be taken, those plants which came into bloom very early or very late the season previous were excluded. It is likely that this is a matter of slight importance. The plants from the falcata portion of the plat were cut June 19 and 20 and immediately weighed on a scale sensitive to 5 grams. The sativa portion of the plat was cut June 24 and 25 and similarly weighed. It is not apparent that this difference in time of cutting had any bad effect upon the experiment. 3 The comparisons in this experiment, properly speaking, were not between hybrids and their parental forms, but rather between hybrids and offspring of the parents breeding true to the respective specific characters, sativa or falcata. Comparison was really made between hybrid plants and their half-sibs, as the pistillate parent was known in all cases. WALDROX I FIRST GENERATION ALFALFA CROSSES. 135 In arranging in series the weights were grouped into classes of 100 grams. The distribution of the frequencies of the four groups is shown in Table I. Table i. — Distribution of frequencies of weight in grams of alfalfa plants, Medicago sativa, M. faleata, and their reciprocal hybrids. Plant group. N CO ro -1- .mvO C--00 00 •0001 -106 IOOI- IIOO. hi M <N u \r> vo 160 1- 1700. y Total. M. sativa 2. 6 24 42 68 68 50 36 17 5 3 I 322 M. sativa X faleata hybrids 1 I 6 5 5 19 17 15 II 20 8 4 4 4 I 1 122 M. faleata 3 28 50 7i 57 50 29 28 9 2 I 328 M. faleata X sativa hybrids 6 13 17 24 32 41 23 25 27 12 6 2 3 1 232 The most striking fact about these distributions is the greater ranges of the hybrid plants compared with those of their pistillate parents. This indicates a greater amount of variation among the hybrids than among the parents. In Table 2 are shown the means, standard deviations, and coeffi- cients of variability, with their probable errors. Table 2. — Variation constants of weight of plants of Medicago sativa, M. faleata, and reciprocal hybrids. Plant group. Mean. Standard deviation. Coefficient of variability. M. sativa M . sativa X faleata hybrids M . faleata M. faleata X sativa hybrids 636.84 ± 7.O4 964.43 ± 17-78 543-49 ± 7-17 776.36 ± 11.68 187.33 ± 4-98 291.21 ± 12.57 192.65 ± 5-07 263.80 ± 8.26 27.27 db O.78 30.20 ± I.42 35-45 ± 1-04 33-98 ± 1. 18 In Table 2, one notes first that the weights of the M. sativa plants are significantly greater than the weights of the M. faleata plants. This is probably about what one would ordinarily look for, judging by the comparative amount of growth made by the two species. Data secured by A. C. Dillman at Newell, S. Dak., and at Akron, Colo., and presented by Oakley and Garver (3) show that in nearly all cases plants of M. sativa grown individually weigh more than simi- larly grown M. faleata plants. One notes again the very much greater weight of the hybrids over either parent. The difference is so decisive and so striking that the probable errors add but little weight. The close agreement in the standard deviations of the M. sativa I36 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. and M. falcata plants is very striking. The difference is about the same as either of the probable errors. The standard deviation of either group of hybrids is far in excess of the parents. The average excess of the two groups of hybrids is 46.2 percent, very nearly the same figure as in the case of the means. In spite of the much greater absolute variability of the hybrids, compared with their parents, the relative variabilities, as expressed by the coefficients, are about the same. This obviously comes about because both the means and standard deviations of the hybrid groups are greater than those representing the parents. The relative varia- bility of the hybrids is intermediate between the two parents. There would seem to be no a priori reason why the hybrid plants should not be more variable than the parents if the latter are in a heterozygous condition, as was obviously the case in this experiment. /J Z7 20 36 73 JZ FAMILIES STUDIED Z3 8 Fig. 10. Weights of plants of Medicago sativa, lower line, and hybrids, upper line, from M. sativa as a pistillate parent. As previously stated, the offspring of each plant was planted as a unit. In figures 10 and 11 the weights of the hybrid plants are com- pared with the weights of their respective parental forms, the latter being plants of M. sativa in figure 10, and of M. falcata in figure 11. Each abscissa represents the offspring of a known pistillate parent. In all cases one finds the hybrid plants to be heavier than those of their parent varieties. In figures 12 and 13 are presented frequency polygons of the weights of the plants. The solid line in each case indicates the weights of the M. sativa (or M. falcata) plants and the broken line the corresponding hybrid plants. WALDRON I FIRST GENERATION ALFALFA CROSSES. 137 In figure 12 a theoretical curve of Pearson's Type I (2) has been fitted to the frequencies of weight of plants of M. sativa. This is 57 SZ S3 82. 62 Fig. 11. Weights of plants of M. falcata, lower line, and hybrids, upper line, from M. falcata as a pistillate parent. shown by the dotted line. The formula for this curve as developed was : (X \ 31.96119 / X \ 128.252)1 i H I ( 1 I 11.86526/ V 47.61236/ The origin was taken at the mode, 614.80. The skewness is posi- tive and equal to .1172. This curve gives a very close fit to the observations. This type of curve seems to be common when one is / — /— \ \ . / / —4 \ \ \ / I 1 \ +A- t 1 V \ \ / / / 1 N l \ \ \ / / ( 1 \ \ -i 1 \ \ \ f 1 \ < \ 1 * 1 — v— \ > AO 78 16 ^14 t I- 6 4 Z 150 ISO 550 450 550 650 750 850 350 7050 J1S0 tZSO 1350 145V 1SSO 1L5D 1750 WEIGHT OF PLANTS IN GRAMS Fig. 12. Variation in weights of plants of M. sativa, solid line, and hybrids from M. sativa as a pistillate parent, broken line. The M. sativa weights are fitted with a Type I curve, shown by the dotted line. i3» JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. dealing with the weight of plants or parts of plants. I hope to show other examples of this type in data to be presented later. The poly- gons of the weights of the hybrid plants, in both cases, shown in figures 12 and 13, are quite irregular, with a tendency toward bimo- dality. The fewer number of hybrid plants would account for some of the irregularity. ISO 250 350 +50 550 650 750 850 350 1050 1TS0 7250 1350 MSO 1SSQ WEIGHT OF PLANTS IN GRAMS Fig. 13. Variation in weights of plants of M. falcata, solid line, and hybrids from M. falcata as a pistillate parent, broken line. The mean of the falcata X sativa plants is very decidedly less than the mean for the reciprocal hybrids, the sativa X falcata plants. This difference might be interpreted as being due to differences in soil effects, the two groups occupying either half of a long strip of land. However, the soil was very uniform in character. A study of the differences in plant weights between the hybrids and their parental varieties along the course of the plat indicates that the difference in weight was not due to a difference in soil. Considering each parent plant offspring as a unit and securing differences between the means of the hybrids and their parents in lots of five, we have the following series of means shown in Table 3 distributed along the course of the experimental strip of land. waldron: first generation alfalfa crosses. 139 Table 3. — Mean of differences between hybrids and parents. sativa and satn'a X falcata. Number. . 2 3 4 5 6 7 8 9 Means. . . 311.6 342.4 341-6 , 386.4 203.3 246.9 354-3 216.5 219.0 M.falcata and falcatay, sativa. While these figures are not conclusive, they indicate rather plainly that there is some intrinsic difference in the yield differences in the two large groups rather than any marked soil difference. Another possible interpretation is that the reciprocal crosses gave different intrinsic results as regards weights of plants. Such a result is so uncommon that it could not well be accepted without excellent evi- dence. A proper explanation of the disparity in yields as here indi- cated must evidently await further experimental work. LENGTH OF STEMS. A stem was taken at random from most of the plants as they were weighed and laid aside to be measured. It would have been a con- siderable task to have selected the longest stem in each case. It is probable that with the considerable number of stems used a reason- ably good sample has been obtained. The variation constants are presented in Table 4. Table 4. — Variation constants of length of stem in centimeters in plants of Medicago sativa, M. falcata, and reciprocal hybrids. Plant group. Mean. Standard deviation. Coefficient of variability. M. sativa 85.54 ± O.67 15.62 ± O.47 18.26 ± O.57 M . sativa X falcata 87.09 ± I.09 18.63 ± -77 21.39 ± -93 M. falcata 80.75 ± .64 14.16 ± .45 17.54 ± -58 M. falcata X saliva 87.26 ± .86 15.53 ± -6l 17.79 ± .72 In this case the stems of M. falcata are shorter than those of M. sativa. This corresponds to what was found in regard to weight of plant, but the difference is less comparatively than in regard to weight. The length of stems of both hybrid groups is greater than their pure parents, but significantly so only in the hybrids of the M. falcata X sativa cross. In regard to the standard deviations, the largest differences are scarcely above the horizon of significance. The standard deviation of the M. sativa X falcata hybrids is greater than that of the other hybrids for both plant weight and stem length. The coefficients of variability do not correspond to those secured from the weight of I4O JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. plants and the differences are not of much importance. It is obvious that the greatly increased weight of the hybrid plants over those of the pure species was due principally either to the increased number of stems or to their increased size, or perhaps to both of these factors. No measurements were made on the diameter of stems. EFFECT OF WINTER INJURY. As previously indicated in this paper, notes were taken as to the spring condition of stand. Plants were given markings from o to 10, the value 1 being given to the weakest plants. These data can not be considered of very great value, as quantitative characters were secured by judging rather than by weighing or measuring. This statement of course does not hold true as to the absolute amount of winterkilling. The distribution of the data is shown in Table 5. Table 5. — Spring condition of stand in two species of alfalfa and their reciprocal hybrids. Class value. Group. Total. Mean. Coefficient of variability. 2 3 4 6 7 8 9 10 M. saliva. . . . 1.942 150 73 61 105 171 235 350 353 259 125 60 5-53 ± -04 46.99 ± .61 M. saliva X falcata. . . . 168 1 1 1 3 3 13 26 3i 33 16 31 7-0i ± .13 36.56 ± I.50 M. falcata. . . 1. 153 12 10 16 40 118 252 362 205 107 25 6 5.76 zb .02 27.58 ± .24 M. falcata X sativa 789 26 6 3l 8 16 48 87 178 211 117 89 7.27 ± .05 29.IT ± .53 The percentage of plants winterkilled in each group was as follows : M. sativa 772 ± 0.27 M. sativa X falcata 6.55 ± 1.91 M. falcata 1.04 ± 0.30 M. falcata X sativa 3.30 ± 0.64 In Table 5 the difference between the means of the plants of the M. sativa X falcata cross and those of the M. sativa parent is quite significant. As I have indicated, this could have come about quite obviously either because the plants of M. sativa suffered compara- tively more winter injury or because the plants of the M. falcata X sativa cross were larger, due to heterosis, or both. The first named factor may have entered to some extent, but the latter factor is probably most important. Practically the same relation is seen to exist between the two other means shown in Table 4. The mean of M. falcata is quite significantly less than that of the M. falcata WALDRON : FIRST GENERATION ALFALFA CROSSES. 141 X sativa hybrids. In this case, the effect of the winter would tend to offset the heterotic effect of the hybrid plants. Probably another factor enters here quite appreciably. This is the comparatively slow early spring growth of the M. f ale at a plants. When these notes were taken on May 24. the M. falcata plants had probably not yet fully awakened from their former dormant condition. The plants of M. falcata are later in starting spring growth than are the plants of M. sativa. One notes the greater variability of the plants of M. sativa and of the M. sativa X falcata hybrids than of the other group. With regard to the percentage killed, significant differences between the group containing the M. falcata plants and the other groups are to be noted. These M. falcata plants are generally recognized as extremely hardy. Practical Applications. "When such marked increases in weight are obtained from the F x hybrids of the plants of a farm crop, one asks at once if the results can be given practical application in the field. It seems to me that the possibilities in this case must be regarded as rather dubious. The problem resolves itself into two parts, the obtaining of the hybrid seed and the results one may expect in the field after the hybrid seed has been sown. Considering the use of the parents indicated in this article, one remembers that the seed coming from M. sativa as a pistillate parent was hybridized less than 8 percent (4). Seed hybridized only to this extent would scarcely influence the weight of the harvest very appreciably, even if the hybrid plants weighed 50 percent more than the nonhybrid plants. The seed from M. falcata, it is true, was hybridized over 40 percent, but the seed production of this species is so meager that it could not be considered from the standpoint of farm practice. The probability of securing hybrid seed between these two species in sufficient quantity for field sowing seems rather out of the ques- tion. There seems to be no compulsory method of obtaining hybrid alfalfa seed, as there is in the case of maize, where one variety can be detasseled. A greater or less amount of hybridized seed could almost certainly be obtained by growing the F x hybrids alternately with plants of M. sativa. However, we do not know the amount of increased w r eight which would accrue with such crosses. Even presuming that hybridized seed were obtainable, it would be unsafe to reason from the results presented here that the yields under 142 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. farm practice would be much increased. The increased plant weights resulted evidently thru an increase in number of stems and it seems evident that this factor could be controlled in considerable measure by differences in rate of seeding. However, if a thin stand were present for any reason, the hybrid plants would evidently be given an opportunity to show an increased weight per plant. The yield per acre would thus be increased. In this* connection it should be con- sidered that when alfalfa is grown where water is the limiting factor, as under semiarid conditions, a thin stand insures the greatest success. In such a case one might suppose the hybrid plants would have the larger number of stems per plant, when compared with nonhybrid plants. But when the water available to the plant becomes the lim- iting factor, one wonders if the tendency for an increase, either in length of stem or in number of stems per plant, would be of much avail toward increasing the yield, except perhaps on certain occasions when the water supply was much above the average. In such a case we would have the machinery present for high yields, ready to func- tion in periods of abnormal water supply. When hybridized seed is once obtained, the resulting crop, it should be remembered, would retain its increased vigor during the life of the plants constituting the crop. One must consider the possibility of the hybrid plants making a more economical use of water than the parents. I do not know that this point has been investigated for alfalfa, but Briggs and Shantz (i) made eight determinations of the water requirement of maize hybrids in comparison with their parents. The yield of dry matter of the hybrids was notably in excess of that produced by the parents, but the water requirements of the hybrids was essentially the same as that of the parents. Summary and Conclusions. This paper deals with weight of plants of the first generation hy- brids secured from crossing the two species, common alfalfa, Medi- cago sativa, and yellow-flowered alfalfa, M. falcata. The hybrids showed a much greater weight per plant than either the M. sativa or M. falcata plants grown under similar conditions. This increase in weight was 47.5 percent. The absolute variability of the hybrids was much greater than of the parents. Data on length of stems were taken of the hybrid and the non- hybrid plants. Significant differences were not shown. The in- creased weight per plant was then probably due to the increased num- ber of stems per plant. WALDROX : FIRST GENERATION ALFALFA CROSSES. H3 Winter injury was comparatively slight, but the plants of M. falcata showed significantly less killing than the other groups. While such an effect of increased vigor thru hybridization might be put to profitable field practice in such a crop as maize, it is not evident that this can be done with alfalfa because of the difficulty of securing the hybridized seed. Literature Cited. 1. Briggs, L. J., and Shantz, H. L. Influence of hybridization and cross- pollination on the water requirement of plants. In Jour. Agr. Re- search, 4 : 391-402. 1915. 2. Eldertox, W. Palix. Frequency Curves and Correlation, p. 53. London, 1906. 3. Oakley, R. A., and Carver, Samuel. Mcdicago falcata, a yellow-flowered alfalfa. U. S. Dept. Agr. Bui. 428. 191 7. 4. Waldrox, L. R. Cross-fertilization in alfalfa. In Jour. Amer. Soc. Agron., 11 : 259-266. 1919. 5. Westgate, J. M. Variegated alfalfa. U. S. Dept. Agr., Bur. Plant Indus. Bui. 169. 1910. TEMPORARY ROOTS OF THE SORGHUMS. 1 John B. Sieglinger. Determinations of the number of temporary roots 2 and secondary rootlets 3 have been made for the common cereals, wheat, oats, barley, and corn. In connection with an investigation dealing with root development of the sorghums, it was found desirable to learn the number of temporary roots of certain varieties of this plant. The varieties used were Sunrise kafir, Dwarf milo, feterita, Manchu kao- liang, and Acme broomcorn. The studies reported were conducted in the greenhouse at the Woodward Field Station, Woodward, Okla., during the winter of 191 9-1 920. A preliminary experiment consisted of germinating 100 kernels of each of the five varieties of sorghum between blotting paper. After 13 days the germinated seed were counted and the seedlings examined 1 Contribution from the Office of Cereal Investigations, Bureau of Plant In- dustry, United States Department of Agriculture, Washington, D. C. Re- ceived for publication February 17, 1920. 2 Wiggins, Roy G. The number of temporary roots in cereals. In Jour. Amer. Soc. Agron., v. 8, no. 1, p. 31-37. 1916. 3 Walworth, E. H., and Smith, L. H. Variations in the development of sec- ondary rootlets in cereals. In Jour. Amer. Soc. Agron., v. 10, no. 1, p. 32-35. 1918. 144 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. to determine the number of temporary roots. The results obtained are presented in Table I. Table i. — Germination of five varieties of sorghum and number of temporary roots on each seedling. Variety. C. I. No. Number of kernels sown. Number germinated. Number of temporary roots per seedling. 472 IOO 84 I Dwarf milo 332 IOO 92 I Feterita 182 IOO 99 I Manchu kaoliang 171 IOO 86 I Acme broomcorn 243 IOO 84 I As noted in Table 1, only one temporary root, the radicle, devel- oped on any seedling. Five hundred kernels of each of the five varieties of sorghum were then sown in a greenhouse bed of screened river sand. The seed germinated well and the plants apparently made normal growth. Plants were dug and the roots washed out. Over 100 seedlings of each variety were examined at different times and in every case the radicle was the only temporary root present. Later, one hundred kernels each of the five varieties of sorghum and a like number of kernels of Kanred wheat and Mexican June corn were sown in a bed of sand. On the ninth day after seeding more than half of each row was dug and the roots washed out and examined. The results obtained are shown in Table 2. Table 2. — Seedlings of wheat, com, and various sorghums grouped according to the number of temporary roots, the total number of seedlings and temporary roots, and the mean number of temporary roots to the seedling. Variety. Numbers of seedlings falling into classes having following numbers of temporary roots. Total num- ber of seedlings examined. Total num- ber of temporary roots. Mean num- ber of temporary roots. 58 219 3-77 6l 270 4.42 42 42 1. 00 63 63 1. 00 41 41 1. 00 46 46 1,00 34 34 1. 00 Kanred wheat .... Mexican June corn Sunrise kafir Dwarf milo Feterita Manchu kaoliang. . Acme broomcorn . . The results shown in Table 2 indicate that sorghums of the milo- durra, kafir, kaoliang, and broomcorn groups have but one temporary sieglinger: temporary roots of sorghums. 145 root, the radicle. These results were obtained under conditions in which Kanred wheat produced a mean of 3.77 temporary roots to the plant and Mexican June corn a mean of 4.42 temporary roots to the plant (radicle included). In sorghums, as with other cereals, the radicle appears first, then the plumule. The radicle develops several branch roots. In sand cultures, three or four days after the plumule appears at the surface, the first node is distinctly developed at about a half inch below the surface of the soil and from this node a whorl of permanent roots is gradually developed. After the permanent roots begin to function, the temporary root soon decays. Summary. Studies of the root development of certain varieties of sorghum under greenhouse conditions have shown that : 1. The radicle is the only temporary root developed in sorghums. 2. Shortly after germination the first node develops below the sur- face, and from this node the first permanent roots develop. AGRONOMIC AFFAIRS. MEMBERSHIP CHANGES. The membership reported in the March issue was 524. Since that number went to press, 13 new members have been added, 2 lapsed members have been reinstated, 7 members have resigned, and 1 has died, making a net gain of 7 and a present membership of 531. The names and addresses of the new and reinstated members, the names of the members resigned or deceased, and such changes of address as have come to the notice of the officers, f ollow : New Members. Barnes, Earl E., Dept. of Soils, O. S. U., Columbus, Ohio. Briggs, Fred N., 122 Hilgard Hall, Berkeley, Cal. Crist, Henry, Farm Crops Section, I. S. C, Ames, Iowa. Dickey, J. B. R., State College, Pa. Gooding, T. H., Agr. Expt. Sta., Lincoln, Nebr. Hurst, L. A., 2001 Sixteenth St. N. W., Washington, D. C. Kondo, M., Ohara Agr. Inst., Kurishiki, Okayama, Japan. Lindstrom, E. W., Dept. of Genetics, U. of W., Madison, Wis. McCool, M. M., Soils Dept., Mich. Agr. College, East Lansing, Mich. I46 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Owens, J. S., 426 E. College Ave., State College, Pa. Schertz, F. M., 1305 Farragut St. N. W., Washington, D. C. Taylor, J. W., 11 16 Tenth St. N. W., Washington, D. C. Trace, Carl F., 331 N. 17th St., Manhattan, Kans. Members Reinstated. Hurst, J. B., Medford, Okla. Wood, Casper A., College Station, Texas. Members Resigned. Abell, M. E., Clevenger, C. B., Hall, Thomas D., Bancroft, Ross L., Erdman, Lewis W., Willey, Leroy D. Birchard, F. J., Member Deceased. Baker, John W. Changes of Address. Boyack, Breeze, Thompson Falls, Mont. Clark, Chas. F., Presque Isle, Maine. Davis, L. Vincent, 340 S. Lumkin St., Athens, Ga. Langenbacher, R. A., County Agent, Macon, Mo. McGuffey, C. Carl, 308 E. Lincoln St., Ada, Ohio. Petry, Edward J., 1218 E. Catherine St., Ann Arbor, Mich. Rast, Loy E., 408 Continental Bank Bldg., Shreveport, La. Richards, Phil E., Morganfield, Ky. Stewart, H. W., New Soils Bldg., College of Agr., Madison, Wis. Weeks, Chas. R., State Farm Bureau, Manhattan, Kans. Woodard, John, Hull Botanical Lab., Univ. of Chicago, Chicago, 111. A NEW COMMITTEE ON VARIETAL STANDARDIZATION. On recommendation of the Committee on Varietal Nomenclature at the Chicago meeting last November, the Society voted to disband the old committee and recommended the appointment of a new com- mittee on standardization of varieties. The functions of this com- mittee are stated on page 350 of the December issue of the Journal. The new Committee on Varietal Standardization has just been named by President Harris, as f ollows : R. A. Oakley, U. S. Dept. Agr., Washington, D. C, chairman. J. H. Parker, Agr. Expt. Sta., Manhattan, Kans. H. K. Hayes, University Farm, St. Paul, Minn. E. F. Gaines, Agr. Expt. Sta., Pullman, Wash. H. H. Love, Cornell Univ. Agr. Expt. Sta., Ithaca, N. Y. NOTES AND NEWS. 147 H. S. Hastings, Atlanta, Ga. George Stewart, Agr. Expt. Sta., Logan, Utah. J. Allen Clark, U. S. Dept. Agr., Washington. D. C. A. B. Conner, Agr. Expt. Sta., College Station, Texas. L. H. Smith, Agr. Expt. Sta., Urbana, 111. NOTES AND NEWS. H. H. Biggar, scientific assistant in the Department of Agriculture, who has been engaged in studies of corn root and ear rots at Bloom- ington. 111., for the past year, has resigned, effective April 30, to become associate editor of The Northwest Farmstead and Dakota Fanner, with headquarters at Minneapolis. Breeze Boyack, assistant agronomist at the Colorado station, has resigned to engage in ranching in Montana. A. E. Ewan, superintendent of experimental fields at the Kentucky station, resigned December 13, 1919, and was succeeded by S. C. Jones. T. E. Keitt resigned January 1 as chemist and agronomist of the Georgia station. C. S. Knight, dean of the college of agriculture of the University of Nevada and agronomist of the Nevada station, has resigned, effec- tive June 30, to become secretary of the Reno, Nev., Chamber of Commerce. D. R. Johnson, assistant professor of agronomy at the Oklahoma college and station, resigned February 15 to accept a position at the Iowa college. Ivar Mattson, superintendent of the Tribune (Kans.) Branch Ex- periment Station, has resigned and has been succeeded by G. E. Lowery, formerly agricultural instructor in the high school at Tribune. E. G. Montgomery, professor of farm crops in Cornell University, is on six months' leave of absence and is now in charge of the Foreign Markets Service of the Federal Bureau of Markets. J. W. Nicolson, secretary of the Michigan Crop Improvement Asso- ciation, is now manager of the seed department of the Michigan State Farm Bureau. I48 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Chas. R. Weeks, for the past four years superintendent of the Hays (Kans.) Branch Station, has resigned, effective May 1, to be- come secretary of the Kansas State Farm Bureau, with headquarters at Manhattan. Leroy D. Willey, superintendent of the Sheridan, Wyo., Field Sta- tion, has resigned and has been succeeded by R. S. Towle, formerly of the Northern Great Plains Field Station, Mandan, N. Dak. Meeting of Advisory Board. The Advisory Board representing the American Society of Agron- omy on the Division of Biology and Agriculture of the National Re- search Council met in Washington, D. C, March 19 and 20, and presented matters of interest to agronomists to Chairman McClung. The action of the Council will be reported at a later date. The meet- ing was attended by C. V. Piper, chairman, Dr. J. G. Lipman, Prof. C. A. Mooers, and Prof. L. E. Call. The Division of Biology and Agriculture of the National Research Council is composed of representatives of the American Society of Agronomy, the American Society of Bacteriologists, the Botanical Society of America, the American Genetics Association, the American Society for Horticultural .Science, the American Phytopathological Society, the Society of American Foresters, the Society of American Zoologists, and the Ecological Society of America, with eleven mem- bers at large. The executive committee of the Division consists of Dr. C. E. McClung, professor of zoology in the University of Penn- sylvania, chairman; Dr. L. R. Jones, professor of plant industry in the University of Wisconsin, vice-chairman; I. W. Bailey, professor of forestry at Bussey Institute ; F. R. Lillie, professor of zoology in the University of Chicago ; G. R. Lyman, pathologist in the U. S. Depart- ment of Agriculture ; H. F. Moore, deputy commissioner of the U. S. Bureau of Fisheries; and Dr. A. F. Woods, president of Maryland State College. JOURNAL OF THE American Society of Agronomy Vol. 12. May, 1920. No. 5. GUAM CORN. 1 Glen Briggs. Corn (maize) is one of the principal food crops of the Island of Guam. It is not a native plant, but was introduced from Mexico nearly 250 years ago. During these two and a half centuries it has been grown by a people who do not practice, nor even understand, any method of improving corn. Guam corn is a variety which has been found to be well adapted to the tropical conditions existing in some of the Pacific Islands and particularly in Guam. Locally, it is known as native corn. Until recently, it received no attention in the way of improvement. While no evidence can be brought forward to prove the claim, it is more than probable that the corn, by a process of evolution, has changed considerably from the original introductions, as almost cer- tainly it has been kept largely free from mixtures. Recently, how- ever, among certain of the natives a superstition has developed that a slight yellow mixture will keep away disease and insect pests. Many introductions of corn from the mainland of the United States and from over forty tropical countries have been made at various times since the American occupation of the island. None of these has proved to be so well adapted to tropical conditions nor to local en- vironment as the Guam corn. Only one type of corn is grown on the island, the present crop being estimated as approximately the 486th in the process of evolution. This assertion is based on careful study of the early history of the 1 Contribution from the Guam Agricultural Experiment Station, Agana, Guam. Received for publication February 28, 1920. 149 I5O JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. island. The first mention of corn growing in Guam is found in the annals of the early mission, where it is related by Padre Garcia (2 ; 7, p. 402) 2 that on the night of October 15, 1676, the natives destroyed a field of maize which was the principal sustenance of the missionaries and soldiers. Two years later, it was said that the natives were be- coming fond of maize, altho they did not make bread of it, not having implements for its preparation. In Guam the principal planting season is in April or May, the second crop being planted in November or December. The time of planting depends upon favorable weather conditions. The first crop matures during the season of light rains which precedes the heavy rainy period ; the second crop is grown in the season following the heavy rains and matures during a favorable drying season. About 120 days are required to mature a crop. From these data it is seen that about 243 years have elapsed since corn growing in Guam was first mentioned and, inasmuch as it is and has been a common practice of the Chamorro people (natives of the Mariana Islands) to save the seed of the preceding crop for future planting, the present corn crop is attaining its 486th cycle. Another reason why the corn is nearly always planted from the preceding crop is that the seed rapidly deteriorates in this humid climate, or is subject to attack from weevils which are very prevalent. These facts make it almost im- possible for the farmer to keep his seed an entire year and, until recently, there were no storage facilities for seed. Crozet (4), who visited Guam in September, 1772, gives Tobias, then the Spanish Governor of Guam, credit for having introduced corn into Guam. This is a mistake, as Padre Garcia's record was made nearly 100 years before. However, it is certain that Tobias encouraged the cultivation and use of maize more than former gov- ernors. It remained for Felipe de le Corte, Governor of Guam from 1855 to 1866, to encourage larger plantings and to study means of preserving or storing corn in large quantities from one season to the next. During this time underground granaries were established upon the island. The remains of one with the grinding implements are still to be seen near Inarajan. That there was probably no corn on the island in the early days is shown by the fact that the Dutch expedition (7, p. 13), under Oliver Van Noort, touched at Guam in 1600 on its way from the South American coast to Manila and obtained supplies, but no corn was mentioned among the provisions brought on board. The noted nat- uralist, William Dampier (5), speaks of the crops growing in Guam 2 Reference is to " Literature cited," p. 157. BRIGGS: GUAM CORN. before the eighteenth century, but does not mention corn in his list of several cultivated plants. Later, Anson (i) visited the Mariana Islands and spent some time there, but, altho he speaks of many of the plants growing on the islands, he does not mention corn. While the old records show that corn was growing on the island before both of these two last-mentioned men visited Guam, it surely was not in large areas, nor was it largely used as food by the natives. No doubt the present corn was originally introduced from Mexico by the Jesuit missionaries, all the evidence being strongly in favor of that theory. No earlier records than that of Padre Garcia mention corn. Many of the words commonly used in connection with the means of preparing corn for food are purely Mexican. It is well known that the original home of many plants can be very accurately deter- mined by their vernacular names. Many of the names and uses of a plant are identical, or associated in some way, with the names in com- mon use in the country from which they were obtained. In some cases, the names are slightly modified to conform to the language spoken in the new country. Again, the original spelling of certain Mexican words which are connected with certain practices in the preparation of corn for food has been retained, while in other in- stances the spelling has been changed to conform to the Chamorro language. Corn is used largely in the form of tortillas, a flat thin cake which is rolled into a paste and baked on a griddle. The word " tortilla " is purely a Mexican term, but has been corrupted into " titiyas " to con- form to the Chamorro pronunciation, or inflection. Along with this method came the Mexican instruments, the " metate " and the " mano " used in the preparation of corn. The former is a low-curved grind- ing stone generally raised on three short legs. The latter is a stone rolling-pin or hand grinder, which may be either cylindrical and slightly tapering at each end, or more nearly square with both a coarse and a fine grinding surface. The spelling and pronunciation of these Mexican words have been adopted by the Chamorro as well as their use. The griddle, formerly made of earthenware and now some- times made of iron, is called " comat " by the Chamorro people. The Mexican term is " comal " but, as a final " 1 " is always changed to a " t " by the Chamorro people, the same name is used, tho it is slightly modified to meet the local conditions. Further proof that corn did not come from other countries is evidenced by the fact that foreign words, other than the few Spanish words and the terms and instru-- 152 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. ments used by the Mexicans, do not appear in the Chamorro language. The following are the few words used : " Gugan," a word meaning to shell or shelling; " fugo," meaning to break (in doubling the stalks) ; " gulic " and " pasa," terms meaning rough and fine grinding, respec- tively, the latter also being used to mean the change of the corn to the paste form in making torillas ; and " coco," a Chamorro word meaning to harvest, to reap, or to gather in the harvest. A number of practices used in the cultivation of corn are also evi- dently adopted from the Mexican customs. Corn is broken over, or the stalk doubled just below the ear, when the kernels show some signs of hardening. This simple and effective method leaves the ears with the point hanging downward, prevents water from collecting under the husks where it would cause germination or rotting of the grain, and also hastens the maturity of the crop. In reality, it pre- vents the loss of the crop in this climate of high temperature and frequent rains. This operation, which is still practiced in some parts of southern Mexico, was evidently introduced from that country by soldiers from Mexico who were doing duty in Guam. In Guam, corn, when prepared for the table, is soaked overnight in lime water. ' This process, which in Chamorro is known as "yaca," softens and loosens the husk or hull until it can be readily removed. The corn is then washed in cold water, when the whole outer hull is removed. It is then ground to a paste between the two grinding stones and made into thin cakes, being baked in the same manner as are the Mexican tortillas. Another fact which indicates the introduction of corn from Mexico is that for the first one hundred or more years after the discovery of the islands, all ships from Mexico sailed from the port of Acapulco to the Philippine Islands, and touched only Guam, or the Mariana Islands, on the outward trip. On their homeward trip, they went far to the North in order to avoid the adverse trade winds. The routes followed by these ships probably more largely account for the introduction of corn from Mexico alone than does any other fact. An interesting feature in the evolution of the Guam corn is the entire lack of seed selection generally practiced in the more modern methods of agriculture. For nearly 250 years this corn was planted season after season with no thought given to the selection of the best ears or to those growing on the best stalks in the field. Nevertheless, the corn conforms more nearly to a type as regards kernel, ear, and stalk, than any other variety seen by the writer in the corn belt of the •United States or elsewhere. While methods of improvement are not BRIGGS : GUAM CORN. 153 practiced by the natives, it must be acknowledged that these people do very carefully select the largest and best kernels after the corn is shelled and dried and save these for seed for the next planting. This practice has resulted in a more fixed type of kernel than is commonly found in other varieties ; the fact remains, however, that this corn has almost become a fixed type, due either to a general mass selection, or to evolution, and probably to both. Recently some mixture has occurred from importations, but this has not been great enough to alter the fixed characteristics already established thru the 250 years of growing one kind of corn. While different writers have varied greatly in their descriptions of the Guam corn, all have agreed that there has been and is only one variety commonly grown on the island. Crozet (4) says : " The cul- tivation of maize especially gives incredible results. It is common to find on the maize fields plants 12 feet high, with eight or ten cobs, 9 to 10 inches long, well stocked with good nourishing grain." Safford (7, p. 402) says that " only one variety of maize is successfully grown on the islands. It is soft-grained and white, resembling that which is most common in Mexico." Thompson (9) has stated that "prior to introductions recently made by this station there was but a single variety on the islands, a hard flinty, white corn with broad, shallow grains and a large white cob." Hartenbower (6), in a picture of the type of corn used in improvement experiments at the Guam Experi- ment Station, shows it to have a large and apparently soft kernel in which a small dimple can be seen. He also states that normally the stalks are relatively low growing, about 5 or 6 feet in height, and most stalks bear two small ears. The shanks are large and, as with most other crops grown in the Tropics, there is a large amount of foliage. There is no uniformity in size or shape of ear, an outcome of a lack of selection or other improvement. In size the ears would be classed as decidedly small, perhaps averaging 5 inches in length, while in shape there is little taper from butt to tip. Both grain and cob are normally white, showing little variation in color. The space between the kernels on the cob is large, and this, with shallow kernels and large cobs, makes a low percentage of grain to cob. Costenoble (3) states that "ordinarily this variety produced ears of fair size, but with a single ear to the stalk. It is seldom that .1 stalk is found bearing two or more ears." He was making selections with a view to developing a type which would have two or more ears to a stalk. It seems hardly possible that this could be accomplished in the short time elapsing between his report and that of Harten- bower, or for the corn to again undergo so complete a change within 154 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. the next three or four years. At present the corn is not of either type, but has a larger percentage of single-ear stalks than it has with two ears to a stalk. This fact seems to be true of all the fields, or at least of a large number visited by the writer. In regard to Crozet's description many of his representations are overdrawn and for that reason should be largely discredited. As to the difference noted between flint and soft corn, it would appear that the writers just mentioned failed to investigate fully the corn on the island, for both types exist, the soft corn occurring in much the larger quantity. This corn, which is more easily prepared for food, is preferred by the native people, who select it for seed. In all probability it has been and still is the type most commonly planted. Probably the difference in descriptions quoted is due to the lack of any definite data upon which to base conclusions. With a view to obtaining accurate information, the writer made some investi- gations and gathered considerable data on the Guam corn which seems to be one of the best tropical corns yet developed in the Mariana and Hawaiian Islands (8). It has shown much promise in the Philippine Islands, and may prove adapted to other similar locations. The stalks of the Guam corn are only medium in size and height, are well proportioned, and bear few, if any, suckers. They gradually taper from the bottom to the top, and bear the ears midway the stalk. The average height of 2,112 nonselected stalks, taken from various fields, was 7.58 feet. The height of the ears from the ground was 3.83 feet, while it was found that the average number of nodes below the ear was 7.4 and above the ear, 4.8. In general, the stalks have approximately 12 leaves each. After selection of this corn thru 10 generations for early maturity, uniformity of type, and for one ear to a stalk, the data show that, in 850 measurements, the average height of the plants was 6.02 feet, the ears were only 2.46 feet from the ground, and the average number of nodes had been reduced only by a half node. While no actual figures are available in regard to the size of the ears in this experiment, ob- servations revealed the fact that they had also decreased in size. Most of the ears are well covered by shucks, which are not quite long enough to furnish ample protection from weevils. In general appearance the ears all seem to be cylindrical ; however, actual data, based on a study of 318 ears which were gathered at random from several fields, showed that 57-86 percent were cylindrical, 26.72 per- cent were slightly tapering, and 15.41 percent could be classed as tapering. The average length of a large number of field-run ears was BRICGS : GUAM CORN. 155 646 inches and the circumference, 6.92 inches. The ears would all- be classed as short, but large in diameter. The cob is large and white with a large shank. Almost without exception the butts are flat, as is shown by the fol- lowing data: 97.48 percent flat; 1.26 percent rounding; 0.94 percent well filled over end; and 0.31 percent with swelled butts. The tips are not well covered and, in some cases, are more or less exposed, showing a considerable amount of protruding cob. This protrusion is probably the greatest weakness of the corn and, to a certain extent, seems to be associated with the long shuck which covers the tips. The results show 98.11 percent of the ears with a protruding cob either to a greater or less extent; 1.26 percent of oval, well-covered tips; and 0.31 percent each of swelled and well-filled tips. The rows on the ears are fairly straight, strongly paired in arrange- ment, and number, as a rule, ten. Sometimes 8, 12, and as many as 14-row ears are found. The 8-row ears are very strongly paired and the space between the pairs is very wide and deep ; the width and depth decrease as the number of rows increase. In an experiment conducted by the station, it was found that the number of rows varied only slightly in two generations due to selection. Table 1 shows the results. Table i. — Effect of selection on rows of kernels to ear. Number of ears harvested from Class. General selection. 8-row selection. 10-row selection. 12-row selection. No. of ears. Percent. No. of ears. Percent. No. of ears. Percent. No. of ears. Percent. Total 706 196 475 34 1 100.00 27.76 67.28 4.82 0.14 280 53 193 33 1 100.00 18.93 68.93 II.78 O.36 365 8l 270 12 2 100.00 22.18 73-97 3-28 •57 416 78 284 49 5 100.00 18.75 68.27 11.78 1.20 With 8 rows. With 10 rows. With 1 2 rows . With 14 rows. The rows are loose on the cob when the corn is well dried. How- ever, it is seldom that the corn is dried on the cob ; as a rule, it is harvested before it reaches maturity and is shelled by hand before the kernels harden. The cobs are large and dry very slowly. On account of the humid weather conditions in Guam it is rather difficult to dry the corn while it is still on the cob. Usually, the germs die and turn black, and yellow and blue molds form between the kernels and the cob. In finding the shelling percentage of the Guam corn, selections were I 56 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. made of 8, 10, and 12-row cars with the following results : 8-row selections, 674. percent ; io-row selections, 68.6 percent; and 12-row selections, 65.8 percent. The last named appeared at first to be con- trary to expectations, but it was discovered that altho the shelling percentage of the 12-row ears was less than the others, the total yield of the corn was greater, being 17.8 percent greater than the 8-row and 1.3 percent greater than the io-row ears. The average weight of 1,378 ears, taken as they were picked from the station field, was 0.5 pound each. The average yield of corn at the Guam Station during 1917 was 37.5 bushels per acre, and during 1918, 37 bushels. While these are not phenomenal yields, they are considered to be fair under tropical conditions. The kernels of the Guam corn are broad, square, with slightly rounded shoulders, and almost parallel on the edges. They have smooth crowns which are slightly dimpled. In general, the kernels are called shallow in depth, but in reality they are medium, being almost a half inch deep. In measuring 1,883 kernels taken from non- selected corn just after it was shelled, it was estimated that it took an average of 2.1 1 kernels to make an inch. The kernels are slightly greater in width than in depth. On the average, 16,003 kernels meas- ured 2.02 kernels to the inch, or a width of almost half an inch to a kernel. In measuring the thickness of kernels, it was found that 2,393 kernels averaged 5.44 kernels to the inch. All were very uni- form in shape and size, the only variation being in those kernels near the butts or tips. In weighing a large number of selected kernels for planting, it was found that 23,856 weighed 8,814.8 grams, or an average of 36.95 grams per 100 kernels. In a lot of field-run seed it was found that 28,918 kernels weighed 10,844.8 grams, or an average or 37.5 grams for each 100 kernels, showing that the kernels must all be about the same weight. In a varietal experiment conducted in 191 2 at the Guam Experi- ment Station, 74 varieties from over 40 tropical countries were under test. Thompson (9) says, the corn grown in this test represented a wide variety of types, grading from the small-grained, flinty, variegated corns from India, Ceylon, Burma, and Formosa, to the large-grained, soft floury mummy corn from Ecuador and Colombia. These two groups, representing the extremes with regard to hard- ness of grain, are also most widely variant in size of kernels, the group from southern Asia requiring from 200 to 220 grains to weigh an ounce, while a variety from Ecuador required 55 grains to constitute an equal weight. Among these varieties only two were promising, one, No. 576, from BRICGS : GUAM CORN. 157 the island of St. Vincent, and the other, No. 589, a very similar variety obtained from St. Lucia. These were earlier than the Guam corn, but otherwise had no advantage over it. Later this seed was lost. The Guam corn has proved in a large number of tests to be superior to other varieties when grown in Guam, and in the Hawaiian and Philippine Islands, and for this reason should be grown more widely than it is at present. Literature Cited. 1. Axsox. G. A Voyage Round the World, p. 304-344. London, 1748. Later published in "The English Circumnavigators," by Wm. P. Nimmo. 2. Corte, . History of the Mariana Islands. (Old Church manuscript in Spanish, filed in Guam Agr. Expt. Sta. Library.) 3. Costexoble. H. L. V. Agriculture in Guam. In U. S. Dept. Agr., Off. Expt. Sta. Rpt. for 1907, p. 409. 1908. 4. Crozet, . Voyage to Tasmania, New Zealand, the Ladrone Islands, and the Philippines in the Years 1771, 1772. Trans, by H. Ling Roth, p. 92. London. 1891. 5. Dampier, W. A New Voyage Around the World, v. 1, 5th ed., corrected, p. 291-303. London, 1703. 6. Hartexbower, A. C. Report of the Agronomist in Charge. In Rpt. Guam Agr. Expt. Sta. for 1915, p. 16, 17. 1916. 7. Safford, W. E. Useful plants of Guam. U. S. Natl. Museum, Contrib. U. S. Natl. Herbarium, v. 9. 1905. 8. Sahr, C. A. Report of the agronomy division. In Rpt. Hawaii Agr. Expt. Sta. for 1918, p. 46. 1919. 9. Thompsox, J. B. Corn growing in Guam. In Ann. Rpt. Guam. Agr. Expt. Sta. for 1912, p. 22, 23. 1913. I58 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. THE UNRELIABILITY OF SHORT-TIME EXPERIMENTS. 1 F. S. Harris and N. I. Butt. Young experimenters are likely to be very enthusiastic over their discoveries. After obtaining results covering a year or two, the temp- tation is to rush into print and proclaim to all the world findings that seem perfectly clear. If restraint is exercised, the chances are that what seemed to be trustworthy results are completely upset by some later year. Discoveries that seemed so certain are spoiled by irregu- larities brought in by additional investigations. It is very easy to formulate a law on the first set of data, but the trouble is that the law is likely to cease operation when put to later tests. Unusual results that are likely to lead to early fame offer a particu- larly strong temptation for hasty publication. Glancing over the more important results of agricultural experimentation, however, the list is singularly free from revolutionary discoveries. Most of the sensations have been caused by the presentation of results covering only a short time. A seemingly perfect correlation is frequently found because the number of variates is so few that all conditions are not represented. The complications in ordinary agriculture are so great that precision can scarcely be expected. It is not at all un- common to find variations of 100 percent in the average yields of crops ; sometimes several years in succession will show an average of nearly this much above the average of a long period. Were it pos- sible to eliminate experimental error, these variations would make very little difference in many comparative experiments, but experi- mental error will creep into even the most careful trials. A small error in an exceptional year may affect the results of the experiment for several succeeding years. Treatments affecting the moisture con- dition- of the soil are especially subject to fluctuations caused by years in which the climatic conditions are widely different from the normal. Experimental Results. In order to show the extent of fluctuations in yields during differ- ent years, the influence of these fluctuations on results variously sum- marized, and the specific dependence that should be given to data 1 Contribution from Utah Agricultural Experiment Station, Logan, Utah. Received for publication February 25, 1920. iiarris & butt: short-time experiments. 159 obtained in a given length of time, a number of experiments con- ducted by the Utah Agricultural Experiment Station have been studied from several different angles. It is thought the results are of general interest in helping to stabilize experimental material. The experiments were, for the most part, carried on at the Green- ville Experiment Farm near Logan, Utah. The soil is a deep, sandy loam, rich in lime. The detailed composition of the soil is shown in Utah Agricultural Experiment Station Bulletin No. 115. To make comparisons, the average results of each test for the period were found and the variation in percentage of each year's re- sults from the average determined. This reduces the data to a com- mon basis so that comparisons may be easily made. FLUCTUATIONS IN IRRIGATION EXPERIMENTS. In Table 1 is shown the variation from the average yield of sugar beets in a typical irrigation experiment covering the years 1904 to 1919, inclusive. While the results in the different tests are not strictly comparable in every case because the years considered are not the same, the number of years averaged is great enough to eliminate serious error. Table i. — Percentage variations from the average yield of sugar beets as affected by different quantities of irrigation water. Year. Inches of water applied. 5 15 20 1904 -41 - 9 -18 1906 24 28 1907 24 19 16 2 11 6 1909 15 24 — 12 -24 -18 - 8 -15 - 9 ipil — 8 — 14 10 1912 I — 1 8 — 4 — 18 *9i3 23 53 24 22 23 1914 27 11 24 35 22 1915 — 22 — 11 -38 -17 - 9 1916 - 7 - 3 — 12 - 3 -15 1917 - 5 2 - 6 — 2 -19 1918 28 13 7 14 -13 1919 -61 -47 — 20 -32 ' Average variation 21 21 IS 17 15 Average yield, tons 12.4 15-0 19-5 21.0 230 From Table 1 it is seen that while on the average the crops receiv- ing least water varied most there is a tendency for the variations with I0O JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. the different irrigation treatments to be in the same direction and somewhat the same during the same year. In some years, as in 191 3 and 1 91 8, there is very close correspondence of most of the variations. In others, such as 191 5 and 1919, there is considerable divergence, altho they are in the same direction. It will be noted that in years of wide variation between the treatments, the same treatment does not always hold the same relative position in different years. For in- stance, the 10-inch irrigation varied most from the average in 191 5 whereas in 1919 it varied least of the treatments. Conclusions drawn from the experiment for a short period which included only one of these years of wide divergencies, especially if it came in a year 0: exceptionally large yields as in 1914, might be considerably different from what it would be with sufficient time to include years with an opposite divergence. Data regarding the yield of ear corn in an irrigation experiment are presented in Table 2. The object of giving this material here in connection with Table 1 is largely to show that the deviations be- tween the treatments for different crops are not the same for the same years. Irregularities which might interfere with the correct inter- pretation of the results are seen here as with sugar beets. Table 2. — Percentage variation from the 9-year average yield of ear corn as affected by different quantities of irrigation water. Inches of water applied. Year. 5 10 20 3° 40 — 28 -29 -17 — 12 — 8 6 1912 19 6 - 5 - 7 4 3 1913 28 20 32 22 11 24 1914 - 4 9 — 1 1 - 6 — 1 2 1915 — 20 — 20 - 4 1 — 20 1916 3 3 - 5 — 3 - 8 1917 16 8 23 5 10 9 1918 21 36 20 23 -15 11 1919 • -34 -35 -37 — 22 -28 -26 Average variation. . 19 18 16 11 9 12 Average yield (bu.) 75-5 87.7 83.9 88.7 89.9 82.8 From the two irrigation experiments just cited we may conclude that the variation in yields of the different treatments from their averages are very nearly alike for the same crop during the same year. Therefore, much of the rest of the material of this paper will be presented with only one or two typical moisture treatments in order that the data will not be too voluminous. HARRIS & butt: short-time experiments. 161 FLUCTUATIONS IN MANURING TESTS. In Table 3 are shown results secured in manuring tests with corn where no irrigation water was added and with a 20-inch annual appli- cation, the plats being duplicated in each case. Variations between the different manurings are 4 somewhat wider than between the differ- ent irrigation treatments just discussed. Moreover, the years do not show a consistent relative variation between the manurings, the 15-ton manuring showing greatest fluctuations one year and least another. The average variation shows that the 15-ton manuring does not hold so close to the average as the unmanured or the one manured at the rate of 5 tons and that the average variation with the heavy manuring is proportionately greater when not irrigated than when irrigated. Table 3. — Percentage variations from the 9-year average yield of ear corn as affected by different quantities of manure. Year. Unirrigated. 20 -inch irrigation. Tons manure applied per acre. Tons manure applied per acre. 5 15 5 15 1911 — 10 — 28 — 21 -14 — 12 7 1912 -14 19 26 - 4 - 7 - 3 1913 24 28 39 7 22 26 1914 -II — 4 19 - 4 - 6 - 8 1915 -24 — 20 -34 -21 1 - 8 1916 3 3 - 3 - 9 - 5 - 6 I9I7- 12 16 12 15 5 4 1918 43 21 25 28 23 27 1919 — 25 -34 -62 2 — 22 -39 Average variation . . . 18 19 27 12 11 14 Average yield (bu.) . . 54-7 73-5 72.9 73-i 88.7 95-3 With corn, as with other crops, there are cycles of two or three years of comparatively high yields followed by similar cycles of low yields and years of large variations between treatments followed by years of small variations. These conditions make it hazardous to draw conclusions until the trend of these variations is known. With favorable years such as 1916 and 191 7, two or three years might be sufficient to get results very near the truth; but should the year 1918 in the above experiment be added, the conclusions both relatively and quantitatively would be erroneous when compared with the average over the longer period. 1 62 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. FLUCTUATIONS WITH SEVERE CROPPING. With the unfavorable climatic conditions encountered when farm- ing without irrigation in arid regions, the yields often vary from almost nothing in dry years to 30 or 40 bushels of wheat to the acre in favorable years. Table 4 shows the variations in yield of unirri- gated wheat when continuously cropped, when cropped every other year, and when cropped one year in three. More moisture is avail- able to the crops under the latter two cropping systems than with continuous cropping. Table 4. — Percentage variation from the 11-year average yield of zvheat on unirrigated plats continuously cropped, cropped during alternate years only, and cropped one year in three on the Nephi Experimental Dry Farm. Year. Variation from average yield of wheat when cropped Continuously. In alternate years. One year in three. 1909 42 -8 7 -83 I9IO -24 -49 -75 -45 44 34 1912 -42 -75 -47 1913 -56 -91 -45 1914 133 in 120 1915 24 92 90 1916 — 20 — 21 8 1917 60 61 24 I9I8 11 5 -15 1919 -81 12 5 Average variation 49 59 50 Average yield (bu.) 10.3 19.4 20.2 It will be noted that the first five years run somewhat contrary to expectations with regard to variations from the average, the continu- ous cropping varying independent of the other two treatments in 1909 and 191 1. The conditions in 1909 were abnormally favorable for the plat cropped continuously, while they were exceptionally unfavorable for the other two plats. The spring rainfall was very low during 191 1, so that the continuously cropped plat suffered, while those which had been fallowed the previous year held enough stored moisture to produce crops better than during average years. The same is true of the year 1919. Conclusions drawn from the yields during the first five years are proportionately too favorable for continuous cropping and cropping one year in three. Results from this period also show yields much below what now appears to be the average yield. HARRIS & butt: short-time experiments. 163 FLUCTUATION IN EXPERIMENTS REQUIRING PERSONAL JUDGMENT. The variation in yield of potatoes shown in Table 5 indicates the great error to which experiments requiring the judgment of different people are subject. Potatoes were separated into piles considered marketable or unmarketable. The yields for the years preceding 1910 were much larger than in later years. This helps to reduce the indi- viduality of the crop during the last years or to make them tend toward the larger yields. During the year 1906 when the production was over ico percent above the average there seems to have been a serious error in classifying the potatoes in the unmarketable class for the 15-inch irrigation according to later standards or else there was an unaccountable proportion of marketable potatoes. Several years are required to smooth out a single large mistake in judgment. This great variation may account for the yields of the unmarketable pota- toes during the last years being above the average, while the total and large potatoes are considerably below the average. Table 5. — Percentage variation from the average yield of total, marketable, and unmarketable potatoes without irrigation and with a seasonal irrigation of 15 inches of water. Unirrigated yield. 15-inch irrigation yield. Year. Total. Market- Unmarket- Total. Market- Unmarket- able. able. able. able. 28 47 -59 1903 44 42 18 1905 35 13 53 1906 124 143 32 64 84 -45 1907 48 17 82 21 6 5i 1908 35 50 — 30 24 28 -18 1909 10 — 24 43 53 -25 I9IO - 3 l - 55 3 -23 -45 46 I9II - 42 - 55 -26 -3i -46 7 1912 - 34 - 37 -44 -61 -78 - 6 1913 - 16 — 17 -34 -21 -33 10 1914 - 46 - 65 — 11 -36 -54 19 1915 - 80 -52 Average variation .... 45 46 34 37 47 28 Average yield (bu.) . . 124.16 91.96 42.15 182.04 150.56 40.12 FLUCTUATIONS IN DIFFERENT CROPS COMPARED. The variation of several crops from their average unmanured 15- inch-irrigation yield is shown in Table 6 in order that comparisons of their relative stability during different years might be made. Disre- garding the years when data were not available, the average variation is greatest for potatoes, followed in order by sugar beets, alfalfa, corn, 164 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. oats, and wheat. Yet the highest single variation was with oats in 1908, when it was 75 percent above its 14-year average yield. During several other individual years the variations of the crops were con- trary to the averages. In general, however, the grains might be ex- pected to give reliable results with fewer years than crops such as potatoes and beets which produce higher total yields per acre. Table 6. — Percentage variation from the average yield of wheat, oats, corn, alfalfa, sugar beets, and potatoes, irrigated at the rate of 75 inches of water a year. Year. Variation from average yield of Wrjeat. Oats Com /Yiiaita . Sugar beets Potatoes. 1902 7 3 54 -54 32 1903 9 23 52 — 21 -59 49 1904 16 1905 13 1906 28 11 70 1907 10 26 1908 75 - 9 -57 29 1909 10 — 20 -19 48 - 9 -23 — 33 -18 18 -44 - 7 — 9 - 7 28 30 -46 1912 28 18 11 30 20 — 20 1913 — 11 - 4 10 27 -19 ~ 3 1914 — 20 - 4 — 12 33 69 -33 1915 - 8 - 4 - 3 33 3 -46 1916 -19 — 22 -24 -30 -29 -38 1917 -28 — 22 -38 — 1 - 7 — 2 — 20 38 14 20 — 17 1919 — 9 — 20 -30 — 20 2 1 Average variation . 14.0 18.4 23-5 25-4 27.0 31.8 Average yield 41.95 bu. 67.62 bu. 71.59 bu. 4.23 T. 6.59 T. 175-85 bu. PROGRESSIVE AVERAGE RESULTS. It has become almost a universal practice to consider only the average results of experiments even when there are only two or three years' data. It is thought that a study of averages for each year of a few experiments will prove instructive in showing how completely one may be mistaken by promiscuously making close comparisons with data covering only a short period. Unstability of First Years. — The average yields of grain and straw at the end of each year in tests with wheat at the Nephi, Utah, Ex- perimental Dry-Farm are shown in Table 7. On land cropped two years in three, it was desired to find the difference in yield between land cropped the previous year and that previously fallow. During the first five years the yields were exceptionally low and only during HARRIS & butt: short-time experiments. 165 the years 191 1 and 1912 did the yields of the crop following fallow exceed that of the continuously cropped plats. During the sixth year the yields were high but unaccountably in favor of the plats cropped the previous year. There was practically no difference in average yield of grain between the two plats at the end of the first 5 years. The average at the end of the sixth year, however, was over a bushel to the acre in favor of continuous cropping. From the sixth crop to the present, conclusions opposed to those of the sixth have rather de- cidedly held, nearly every year showing a greater relative difference. The abnormal years at the beginning of the experiment made it im- possible during the first 7 years to arrive at what seems the correct relationship between these two cropping methods. The results are seen still to be erratic both in yield of grain and straw, indicating that close comparisons would not be warranted even at the end of 11 years. Table 7. — Progressive average yields of wheat and straw before and after a fallow year in an experiment with dry-land plats cropped to wheat tzvo years in three, for the period 1909 to 1919. Year. Progressive average yield of grain. Progressive average yield of straw. Before fallow. After fallow. Before fallow. After fallow. Bushels. Bushels. Pounds. Pounds. 1909 13-4 2.5 820 160 11.9 5-4 825 355 I9II 10.6 11.4 760 773 8.9 10.2 635 703 1913 8-5 8.6 606 616 1914 14.4 13-8 1,060 933 1915 14-3 17.0 1,187 1,203 13-5 17. 1 1,117 1,184 1917 14.2 17.8 1. 154 1,240 I9I8 13-7 17.6 1,007 1,207 1919 14.2 17.9 1. 139 1,183 Progressive Shifting of Conclusions. — An experiment to discover the proper quantity of irrigation water to apply to corn manured at the rate of 5 tons to the acre each year has been run for 9 years. During the first year, as seen in Table 8, the yields were about pro- portional to the quantity of water added. The average of the first two years shows that the 10-inch irrigation became less effective than the 5-inch and the 40-inch less effective than the 30-inch and remained so thruout the experiment. The unreliable factor in this experiment is the relation the 40-inch and the 5-inch irrigations held to the other irrigation treatments as the test progressed. Comparisons made at the end of the fourth year would not have been true at the end of the fifth. The yields with 40 inches decreased from first rank in the be- I 66 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Table 8. — Progressive average yield of ear corn without irrigation and with applications of 5, io, 20, so, and 40 inches each season from 1911 to 1919. Year. Progressive average yield in bushels of ear corn irr gated at the rate of inches. 5 inches. 10 inches. 20 inches. 30 inches. 40 inches. 1911 . . . 52.7 62.5 69.2 77-8 82.8 87.6 1912 . . . 70.3 77-8 74-3 80.1 88.2 86.4 1913 78.I 86.8 86.5 89-5 92.0 91.8 1914 • 76.3 89.1 83.4 88.1 91.3 90.I 1915 •• • 72.8 85.4 82.8 88.4 91.0 85.3 I916 . . . 73-3 86.2 83.0 87.7 9O.4 83.8 1917 •• • 75-0 87.4 85-9 88.5 91.6 84.6 1918 . . . 76.7 91.4 87.8 91. 1 93-1 85.5 1919 ■ • • 73-5 87.7 83.9 88.7 89.9 82.8 ginning to last of the irrigations during the final three years and the 5-inch-irrigation yields tended to improve in comparison with the others. Considering the shifting positions of the averages during the eighth and ninth years, the variations in yield as indicated by the aver- ages are still so great that dependable deductions on the relative value of the different irrigations can hardly be made. LENGTH OF EXPERIMENT AND EXCEPTIONAL YEARS. The percentage of sucrose in sugar beets irrigated at three different rates, together with their progressive averages, are given in Table 9 Table 9. — Yearly percentage of sucrose and progressive average percentage of sucrose in sugar beets irrigated at the rate of 10, 13, and 20 inches of water each season from 1907 to 1919. Year. 1907 1908 1909 I9IO I9II 1912 1913 1914 1915 1916 1917 1918 1919 Yearly percentage of sucrose. to inche 17.23 13.90 IO.26 16.40 15.80 17.04 14.55 14.64 14.52 II.87 15.42 13.69 8-59 15 inche 20 inches. 15.86 13.39 n-57 16.35 15.40 18.41 15.02 14.65 15-41 12.66 16.43 15.09 11.31 16.08 14.97 11.84 15.98 14.88 21.18 15.04 15.60 12.76 14.40 15-91 16.43 Progressive average percentage of sucrose. 10 inches. 17.23 15-57 13- 80 14- 45 14.72 15- 11 15.03 14.98 14-93 14.62 14.69 14.61 14-15 15 inches, 15.86 14.63 I3.6I 14.29 14.51 15.16 15-14 15.08 15.12 14.87 15.OI 15.02 14-73 20 inches. 16.08 15.53 14.30 14.72 14- 75 15- 82 15-71 15-70 15-37 15.27 15-33 15.42 to indicate the possible fluctuations of averages covering several years when an unusual season is encountered. The seasons of 1912 and 1916 throw the averages out of proportion to what would be expected HARRIS & butt: short-time experiments. 167 in the more normal years. The 20-inch irrigation took an abnormally high position in 191 2 and remained noticeably high for the next two or three years, when a counteracting low percentage brought it down. It was following this 1912 abnormal year that the percentage of sucrose for the 10-inch irrigation fell below that of the 15-inch irriga- tion and remained there. During the tenth season, 1916, a subnormal percentage of sucrose occurred. The 15-inch irrigation treatment was affected proportionately more than the other two, throwing the treatment out of the relative position in which it- apparently belongs. Being an average of 4 plats this relative shifting of positions is not thought to be caused by experimental error so much as to peculiarity of the season. Summary. This paper includes data from experiments with potatoes, sugar beets, alfalfa, corn, oats, and wheat. Short-time experiments are especially subject to error where a com- plete cycle of seasonal fluctuations is not included. All treatments in an experiment are not affected relatively the same each season ; the amount of divergence varies in different years. Where variations from the average condition are large, a greater number of years are required for accurate conclusions than where the variations are small. Manuring experiments show wider variations from the average than irrigation experiments. Under dry-farming conditions variations are wider than under irri- gation conditions and small irrigations vary more than where the plant does not suffer for water. In these experiments potatoes varied in yield most, followed in order by sugar beets, alfalfa, corn, oats, and wheat. Experiments requiring personal judgment vary more than where measurements are mechanical. Average results during several abnormal years ran contrary to re- sults in normal years that followed. Progressive seasonal averages showed a gradual shifting of the relative positions of part of the treatments in an irrigation experiment with corn. In experiments where variations are comparatively large, an excep- tional season may seriously affect the average of a 10-year period. 1 68 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. THE COEFFICIENT OF YIELD. 1 Frank A. Spragg. Heredity and environment stand side by side in the production of individual plants and animals. An individual with the best inheri- tance will fail under some environments, but the individual with poor inheritance is not capable of good production even in the best sur- roundings. The problem in interpreting experimental results is not how to eliminate environmental influences, but how to compare varie- ties on a uniform basis. It is unnecessary to outline the development of the use of checks in varietal and other experiments, as its growth was simultaneous at most of the important experiment stations. By 1900, the check idea was fairly well understood and generally used, but even after niner teen years, agronomists differ as widely in their use of checks as the soils vary in the different sections. Some depend entirely upon the average of the results obtained from a large number of small plats. This replication idea is a good one, and may give fair results on uni- form soil, but the check idea is also a good one, and must not be overlooked. The method employed at the Michigan station makes use of both of these principles, incorporating replication and checks into one system. The old addition and subtraction method has been done away with, and a ratio (percentage) method put in its place. The results of the interpretations are called coefficients of yield. The coefficient of yield is the quotient obtained by dividing the yield of a variety by the calculated yield of the standard or check variety, growing on the same plat the same year. This is a ratio that becomes unity when the yield of the variety it represents equals that of the standard. It is greater or less than unity when the yield of the variety it represents is greater or less than the yield of the standard. This method of calculating results was adopted after other methods had failed to meet our conditions. For example, in the early days of varietal testing in Michigan, the contrast between the many poor 1 Contribution from the Michigan Agricultural Experiment Station, East Lansing, Mich. Presented at the twelfth annual meeting of the American So- ciety of Agronomy, Chicago, 111., November 13, 1919. SPRAGG : COEFFICIENT OF YIELD. 169 varieties and the few good ones was so great that even crude methods obtained results. The plats were often a rod wide and usually not over 8 rods long. Few checks were used and the soil varied so be- tween them that large experimental errors were created in inter- preting results. The addition and subtraction method of eliminating environmental influences was also used. Trouble came when dealing with such a problem as alfalfa seed production, where some lots pro- duce almost no seed under good environment. The corrected results would often appear to be less than zero. The absurdity of the results proved that the addition and subtraction method was false. We have improved the old methods in two ways, first, by making the checks more frequent, and second, by making the plat determina- tions more reliable thru the use of the principle of replication. With this arrangement, we feel warranted in assuming that the soil varies continuously between checks. This enables the calculation of the comparative yielding power of the check upon each and every plat in the series. In making use of the principle of replication it was early found that the replicated plats could be sown end to end, and that the many soil types would thus be sampled. Each collective replication is harvested as a unit. A large amount of expensive labor is saved and the reliability of the results is increased, because ten years of experience in handling small plats convinces the writer that when a number of small lots are harvested and thrashed separately a small loss for each lot creates a big mistake. The larger lot can be more easily kept from being mixed, allowing the use of the varietal series as increases of new pedigreed sorts, to be further increased if desirable. 2 As a result, the varietal plat is a long and narrow strip, sampling as many soil types as possible and yet close enough to its neighbor to have approximately the same yielding power. The checks would not be possible by any other arrangement of the replication. Experience teaches that the distance between checks should not be more than one-twentieth of the length of the plats. It is better in some cases if this relationship can be less. Two and not to exceed four varieties are planted between each pair of checks. The number depends upon the shape of the tract available for test work and the amount of avail- 2 The breeding work at the Michigan station has produced a number of new commercial varieties. The Alexander and Worthy oats were distributed in 191 1. College Success, College Wonder, and Wolverine oats have been pro- duced since that time. Several varieties of wheat have been distributed, of which the American Banner and Red Rock are the most noteworthy. The Red Rock wheat and Rosen rye have spread from the Atlantic to the Pacific. 170 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. able seed of each variety. Whenever possible, the varietal series is duplicated for the same reason a chemist duplicates his work. The standard or check variety is as pure and as high-yielding a variety as the station possessed at the time of its selection. The same Standard is maintained from year to year to test the variability of seasons as well as variations between different parts of the same field. It is best if the yield of the standard can be obtained yearly from a general field rather than to depend upon the yield of the standard in the series, where frequent alleys exist. The crop draws from the alleys. However, as the plats are all the same size and shape, the results are comparable and their relationship is reliable. In 191 3, when our department obtained a calculating machine, much time was spent investigating ways and means of getting more accurate results. These investigations led to the discovery that is here called the coefficient of yield. It has been tested in all kinds of connections during the past six years and found superior to any of the old methods of interpreting results. The coefficient of yield has proved valuable to the crops work at the Michigan station and should likewise be found valuable to other investigators. In practice, it has been found best to express this ratio to the nearest fourth decimal place or unity plus four places when the coefficient is greater than unity. The expert statistician would say that this is more accurate than the facts justify. The an- swer is that these are figures in process and not the final results. The desire is that this relation will fully represent the numbers involved, so that the end results will be as accurate as the data. When we come to end results, expressed in bushels per acre, they will be given in bushels and tenths only. In the case of a series of varieties that are tested thruout a series of years, the coefficients of yield that represent these varieties are found to be more reliable values, when comparing the yielding power of the varieties in question, than those given in bushels for a series of varieties that have been run different numbers of years. Dr. T. L. Lyon 3 mentioned this difficulty when considering sources of error in field tests. The main trouble is that a series of values in bushels per acre is a weighted series. The yielding power of the standard variety varies from year to year, depending upon varying climatic conditions. The corrected values (bushels per acre) that the experi- menter obtained for a year's results are based on the yield of the 3 Lyon, T. L. Some experiments to estimate errors in field plat tests. In Proc. Amer. Soc. Agron., 3 (1911), p. 113. 1912. SPRAGG : COEFFICIENT OF YIELD. 171 standard variety for that year. The results of another year are based on the yield of the same standard, but nevertheless on a different value, and thus each yield for a certain variety under test receives a different weight, and the average of such a series is a weighted average. To overcome this difficulty in data that already have been calcu- lated in bushels per acre, the experimenter must divide the corrected yields in bushels for a series of a certain year by the average yield of the standard in bushels for the same year. He would thus obtain the coefficients of yield for those varieties that year, provided he has used the ratio method of correcting results. Let him do this for each year and then average the coefficients of yield representing a variety during the period of years of the test. He would now have removed the weights and obtained an unweighted or normal average for that variety and likewise for each variety of the series. When a series of average coefficients of yield have thusi been obtained, it is usually desirable to transform these into bushels per acre for the purpose of publication. This is done by multiplying each average ratio (coeffi- cient of yield) by the average yield of the standard thruout the years. The experienced experimenter knows that environmental in- fluences are heavy weights. Perhaps he may use the average of a lot of tests on the standard variety over the State to calculate his figure representing the average yield of the standard. If he sees best he can do this, to obtain figures for the other varieties based on greater adaptability. The flexibility of the method, using coefficients of yield* allows this to be done. The preceding discussion refers to the handling of old data that have already been calculated in bushels per acre. Coefficients of yield can be more easily obtained from the original data. The yields in pounds per plat are placed in a column, which is usually marked P (plat yield). The next column contains the check yields (C). The third column gives the corresponding P/C. In the case of a check the value P for that plat is also the C and P/C equals unity. The other Cs are obtained by interpolation between checks. These re- sults can be obtained graphically by placing the data on cross-section paper and drawing straight lines between the dots representing the yields of adjacent checks. This generates a broken line across the series known as the " normal." The points in which the normal in- tersects the plat lines are the values for C referred to above. The reliability of these values is indicated by the nature of the normal. It should not be an abrupt zigzag. The quotient P/C is the coeffi- 1/2 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. cient of yield calculated for each plat and can be placed in a third column. The ratio method, upon which the coefficient of yield is based, con- siders the environmental influences as weights that it undertakes to remove, by eliminating the ratio between the yielding power of a plat and the yielding power of an average plat in terms of a standard variety. This ratio of yielding power is the weight that is removed when we take the quotient of the yield of a plat divided by the calcu- lated yield of the standard variety on the same plat. This quotient is the coefficient of yield. 100 P/C is the same thing in percentage form. PK/C is the corresponding result in bushels or tons per acre, when K is the acre yield of the standard or check variety. Table i. — Data obtained from the 1914 bean variety series at the Michigan station. Register No. Strain. Variety. Yield of plat, pounds. Yield of check, pounds. Coeffi- cient of yield. Calcu- lated yield, bushels per acre. 40000. . 81302 Robust 117. 5 H7-5 1.0000 34-19 40IOO. . 36 Shoesmith 71.0 123-7 •5740 19.62 40200. . 40 Boston 69.O 129.9 •5312 18.16 40300. . 41 Unknown 78.O 136. 1 •5731 19-59 40400. . 42 Cook 90.0 142.3 •6325 21.63 40500. . 81302 Robust 148.5 148.5 1.0000 34-19 40600. . 44 Landis, untreated 72.0 150.0 .4800 16.41 40700. . 44 Landis, inoculated 68.0 I5I-5 .4488 15-34 40800. . 44 Landis + 1 : 300 formaldehyde 67.0 153-0 •4379 14.97 40900. . 44 Landis + 1 : 500 HgCh 7i-5 154-5 .4628 15-82 4IOOO. . 81302 Robust 156.0 156.0 1.0006 1 34-19 4IIOO. . 45 Scully 72.5 154.6 .4690 16.04 41200. . 46 Geismar 39-5 153-2 •2578 8.81 41300. . 2 Red Kidney 50.5 151. 8 •332 7 n.38 41400. . 4 White Kidney 57-0 150.4 •3790 12.96 41500. . 81302 Robust 149.0 149.0 1.0000 34-19 41600. . 81901 Selection from Commercial 79.0 145.0 •5448 18.63 41700. . 82103 Selection from Commercial 73-5 141. •5213 17-83 41800. . 84204 Selection from Commercial 61.5 137-0 •4589 15.69 41900. . Early Buff cowpea 17.0 133-0 .1278 4-37 42000. . 81302 Robust 129.0 129.0 1.0000 34-19 The I9i4 bean varietal series may be taken as an example. There were two rows to each plat, which were thrown together by the puller. The plats were 637 feet long, with 28 inches between the rows. The two rows make the plats 4.67 feet wide. The Robust variety was used as a check. Four varieties were planted between checks. Five times 4.67 equals 23.35 feet. In this case the length of. the plats is 27.3 times the distance between checks. Thus we feel warranted in assuming that the soil varies continuously between checks, enabling spragg: coefficient of yield. 173 the calculation of the cheek yield (C) for every plat of the series. There is an extra plat planted for edge outside the last check and next to the road. Table 1 shows the results of the same series, the calculation of the Cs, P/Cs, and yields in bushels per acre. In this case the average yield of the checks was 140 pounds. The area of a plat was 0.06824 acre. This gives the average yield for the Robust bean of 34.19 bushels per acre. This is the K in the calculations. Multiplying each P C by 34.19 gives the corrected yield of the other varieties of the series. Thus far we have discussed only the simple coefficient of yield. Coefficients can be used to compare varieties from the standpoint of several qualities upon which the plant breeder wishes to base his selections. In wheat, yield and quality go hand in hand, and to these hardiness and stiffness of straw must be added, when they are deter- mining factors. Quality can be expressed as a coefficient of yield by baking a loaf of standard each time a baking test is made. In this case the volume of the loaf for a certain variety will be represented by P. The corresponding volume for the check variety at the same baking is represented by C. Then P/C becomes the coefficient of loaf volume. A series of coefficients of yield and quality may thus be determined for a wheat series. If each pair of coefficients are now multiplied 4 together, a series of compound coefficients is deter- mined. The variety having the highest compound coefficient will possess the best combination of yield and quality. Coefficients of hardiness and stiffness of straw may be separately determined, but we usually find they are already determined in the coefficient of yield, because they are factors directly affecting the yield and do not need to be determined further. In the alfalfa breeding work, three considerations are of prime importance. These are hardiness, average yield of hay, and average yield of seed. The seed is a means to an end. The end is hay, but the variety is of little value to Michigan farmers unless it will endure severe winter conditions. The success of alfalfa as a hay crop in Michigan depends upon the success of a Michigan seed industry. In the alfalfa breeding work, where individual records are kept with all the plants and the progeny rows include checks, coefficients 4 One may think that these coefficients should be averaged. The fact that they should be multiplied together is seen, if we suppose a series that are equal to the standard in yield, P/C — 1 for yield. The varieties should then be com- pared directly as the quality coefficients. iX,P/C does not change it, but half of 1 -j- P/ C reduces the variability of the series. 174 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. may easily be determined for hardiness, yield of hay, and yield of seed. These three coefficients multiplied together for each variety or progeny of the series will give the composite value in one figure. This product will be unity for each of the checks. A progeny giving a compound coefficient greater than unity may be considered superior to the check, three things being considered. The variety suited to Michigan conditions cannot be obtained unless the seed is produced in paying quantities. Thus it is necessary to include the hardiest high seed yielders as well as those that are best when hay production is also considered. Representative strains selected on these two bases are included when each new generation of alfalfa nurseries is set out. An experimenter can compare varieties on the basis of their coeffi- cients, simple or compound, much more easily than by means of their yields and qualities measured by any other units. In the case of the check, the coefficients are unity, and unity can be raised to any power and still be unity. It is very easy for the eye to separate the values that are less than unity from those that are greater than unity. All investigators will agree that the prime object in making yield and quality tests is to find the superior varieties, discard the inferior ones, and increase the best ones for the benefit of the public. Any method that will bring out the truth most forcefully is the one to be used. carrier: history of the silo. 175 THE HISTORY OF THE SILO. 1 Lyman Carrier. The term "silo" is of ancient origin and means a grain pit (2). 2 Hermetically sealed granaries either above or below ground, usually partly below and partly above, were in use in the dry Mediterranean countries long before the Christian era. Varro (4) states : Some farmers have their granaries under ground like caverns, which they call silos as in Cappadocia and Thrace while hither in Spain in the vicinity of Carthage and at Osca pits are used for this purpose the bottoms of which are covered with straw ; and they take care that neither moisture nor air has access to them except when they are opened for use, a wise precaution because where the air does not move the weevil will not hatch. Corn stored in this way is preserved for fifty years and millet indeed for more than a century. The Egyptians built batteries of granaries separate from their other buildings. These were constructed of masonry above ground, were conical in shape, and were filled through an opening near the top (5). The grain was taken out through a door near the base. They were used to store grain in years of plenty for years of scarcity. The ac- cumulation of carbon dioxid from partial fermentation of the grain effectually preserved the remainder. Varro (4, p. 172) says: Those who store their grain in the pits which are called silos should not at- tempt to bring out the grain for some time after the silo has been opened be- cause there is danger of suffocation in entering a recently opened silo. Attempts to introduce this method of storing grain into France early in the nineteenth century copied from Underground silos used in Spain failed because of the porous nature of the soil and seepage of water. This trouble was finally overcome by Doyere in 1855, who suggested building masonry silos lined with sheet iron. The Paris Omnibus Company constructed several silos, some underground and some above, after Doyere's plan which were in use for several years (5). The practice of storing grain in underground pits was not confined to any one country or race of people. Some tribes of American In- 1 Contribution from Office of Forage-Crop Investigations, Bureau of Plant Industry, United States Department of Agriculture, Washington, D. C. Re- ceived for publication February 27, 1920. 2 Numbers in parentheses refer to "Literature cited," p. 181. I76 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. dians made use of this method to store their own corn while they were away on winter hunting expeditions (14). It was practiced more or less perhaps by all nomadic peoples. There does not appear to be much connection, however, between this ancient method of storing grain and modern methods of ensiling forage. About all that the old method furnished was the name " silo " for the structure for the new process, from which were de- rived the terms silage and ensiling. As the secret of success of stor- ing grain in silos was to have it dry when ensiled it does not seem probable that the preserving of green forage by the same method would be the natural outcome. ENSILING GREEN FORAGE. It is impossible to say when or where the practice of preserving green forage in pits or silos originated. The statement has been made many times that the process was known and practiced by the Romans. Not much evidence is apparent to substantiate this state- ment. Cato (4, p. 43) does say: As long as they are available feed green leaves of elm, poplar, oak and fig to your cattle and sheep. Store leaves also to be fed to the sheep before they have withered. As the first reference in modern times to the matter of storing green forage for cattle, that of Prof. John Symonds (11) of the University of Cambridge in 1786, was made from observations in Italy of the process of preserving the leaves of trees in casks and pits, it seems highly probable that the practice as far as that class of forage is concerned comes down from the time of the Romans. A French correspondent to The American Farmer (13) in 1875, referring to the preserving of green forage, states : There is nothing positively new in the idea. Since time immemorial vine leaves have been preserved in a green state in the district of Lyons and which has made the reputation of the famous Mt. Dore cheese. If it had been customary for the Roman farmers to ensile green grains and other forages it seems probable that some of the agricul- tural writers of the time would have mentioned it. Their descrip- tions, however, with the exception noted above, deal with the storing of dry seeds in silos and not with green forages. No matter whethe-r the ancients ensiled green forages or not, the modern practice traces directly to the process of making sour hay in Germany and Hungary. This method was called to the attention of CARRIER : HISTORY OF THE SILO. 177 English-speaking farmers in 1843 by Prof. J. F. W. Johnston (6). It consisted of storing green grass, clover, or vetches in pits 12 feet square and 12 feet deep. Salt at the rate of 1 pound to each hundred- weight of green grass was added and the material was thoroly tramped by five or six men. After the pit was filled it was covered with boards and on these was placed a foot and a half of earth. Each pit held about 5 tons of fresh grass. It was found later that salting was unnecessary. The similarity in the methods of making sauerkraut and silage has frequently led to the suggestion that ensiling green forages was the direct application of that well-known method of preserving cabbage. There is some basis for the belief. Sauerkraut-making was an older process than ensiling, judging from available historical data, and was commonly practiced in the countries where the method of making sour hay originated. Murray's New English Dictionary, defining the word sauerkraut, gives the following quotation under date of 1633 from Hart's Diet of the Diseased : They pickle it (cabbage) up in all high Germany with salt and barberies and so keepe it all the yeere, being commonly the first dish you have served in at table which they call their sawerkraut. The fact that the Germans first used salt in making their sour hay also lends strength to this suggestion. The method of making sour fodder was described by a correspond- ent to the American Agriculturist (1) in 1873 from Albrechtsfeld, Hungary. Instead of using pits, it was customary there to dig trenches 12 feet wide at the top, 6 feet wide at the bottom, 12 feet' deep, and 10 to 20 rods long. The process of filling and covering was similar to that described above. He stated that sour fodder could be stored for a few years without injury. Neither of these descrip- tions seems to have created much interest in the subject at the time it was published. Some French farmers tried the method, but for the same cause which resulted in the failure of the first underground grain pits in France, water seepage, the practice was abandoned. The first recorded attempt to ensile green maize was made in 1861 by Herr Adolph Reihlen (15), a sugar manufacturer, near Stuttgardt, Germany. Herr Reihlen had previously traveled in America and had taken back seed of maize which he was growing largely for soiling purposes. He was undoubtedly familiar with the German method of making sour hay and had adapted the method to the storing of beet tops and beet pulp, both of which are easily preserved as silage. As early frosts often killed the maize crop, Herr Reihlen tried storing it I78 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. in trenches. He was so well pleased with the experiment that he gave his experience in a letter dated April, 1862, which was published in the Wurtemberg Wochenblatt. This was followed by another let- ter by the same gentleman dated September 23, 1865, and published in the same paper. These letters were translated into French by M. Vilmorin-Andrieux and published in 1870 in the Journal d' Agricul- ture Pratique. At the time M. Vilmorin-Andrieux prepared his re- port Herr Reihlen had increased his acreage of maize until he filled silos 15 feet wide at the top and slightly narrower at the bottom, 10 feet deep, and with an aggregate length of over three-fourths of a mile. It is also of historic interest to note that Count Roederer of Bois- Roussel in the Department of the Orne in 1867 began to preserve chopped green maize mixed with cut straw B in silos (5, p. 136). This method w r as described in a letter dated June 18, 1870, published in the Journal d'Agriculture Progressive the following week. The pur- pose of this experiment was to render the straw palatable as well as to preserve the green maize. M. Piret, farm manager for A. Houette at Bleneau, Belgium, in 1868 successfully experimented with the en- siling of chopped maize. He constructed in 1870 two pits of masonry. "They were found equally serviceable to those below ground" (12). The publication of these articles in the French agricultural press in 1870, together with a disastrous drouth which prevailed that year thruout France and ruined the hay crop, caused widespread interest to be taken in the subject. M. Moreul of La Grignonniere in 1870 built an aboveground silo of masonry and filled it with unchopped but salted maize (9). His success induced M. Crevat in 1872 to con- struct three pit silos 26 feet by 10 feet at the top, 22 feet by 6.5 feet at the bottom, and 6.5 feet deep. A number of other attempts were made to ensile forage besides these which are here mentioned. Re- ports of experiences with this process kept appearing from time to time in the agricultural press. Another drouth in 1874 brought this subject prominently before the French farmers. The French Agri- cultural Society offered a prize that year to be awarded in 1876 for the best essay on the subject, " Preserving Green Forage." This action resulted in a great many literary efforts on the part of those who had tried the ensiling method. M. Goffart (3), a gentleman farmer of Burtin in Sologne, was one of the ablest of these writers. He had had considerable experience with the German system of mak- ing "brown hay" and had grown maize for forage for a number of years. He built four silos in 1852 hollowed out of the ground and CARRIER : HISTORY OF THE SILO. 179 plastered with Portland cement. He did not claim success for his method, however, until 1873. His earliest silos were too small to be practicable, each holding only about 2 cubic yards. They were used to store cut and mixed straw and maize for immediate feeding. They prolonged his period of feeding green corn fodder three or four weeks but did not effectually preserve the fodder. Goffart states : In 1873, I had a real success, due mainly to accident. . . . Until this time I had hardly believed that the preservation of green maize for a long time was possible and I had very little confidence. As Goffart had at that time the benefit of the successful experience of Herr Reihlen, Count Roederer, M. Piret, and M. Moreul in pre- serving green maize, he can hardly be credited with being the orig- inator of the process. His comprehensive experiments, his useful writings, and his frequent lectures before agricultural societies, how- ever, well earned the decoration of Legion of Honor which he re- ceived in 1876 and the popular esteem in which he has been held in America as the " Father of Ensilage." THE SILO IN AMERICA. The introduction of the silo in America w r as due directly to the publicity given the subject in France. The first appearance of a description of the French process of pre- serving green forage appears to have been in a series of letters from a French correspondent to the American Farmer published in Balti- more, Md. These letters were dated Paris and signed F. C. They ran through a number of years. This correspondent referred to the matter in a letter in the issue of October, 1874. In the January, 1875, issue he wrote : Quite a revolution is taking place in the agriculture of the south of France by the cultivation of maize for green fodder and its preservation in a green state chopped and mixed with straw in trenches for winter consumption. The subject was referred to again in subsequent letters. It was these accounts which induced Francis Morris, of Oakland Manor, Md., to try the method in 1876 (10). In the American Agriculturist of June, 1875, was an illustrated article entitled "Curing Green Fodder" which was translated from the Journal d'Agriculture Pratique. The terms silo was used in this article, probably for the first time in an English publication. In England the terms trenching, pitting, and potting were used. The experiences of M. Piret and M. Goffart were given. Practically 180 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. everything connected with the French method was given in this ar- ticle. This information was restated many times during the next five years. A more complete resume of the subject was published in 1876 in the Annual Report of the United States Commissioner of Agriculture for 1875 (15). This was similar to the article in the American Agriculturist, being copied from the same French journal. The Cultivator and Country Gentleman in its issue of October 21, 1875, had a communication signed B. F. J. (probably B. F. Johnson), Champaign County, 111., from which is quoted the following: Prof. Miles of the Illinois Industrial University has made pretty liberal ex- periments in the ensilage of maize and broom corn seed in the course of this autumn, the outcome of which will be given to the public as soon as the success or want of it in the undertaking has been determined. . . . To ensilage is to bury in silo or pits. This communication is of special interest not only because it records the first American effort to preserve green corn but it also introduces the word " ensilage " into our language. Prof. Manly Miles in the issue of October 5, 1876, of the same paper gave a summary of the experiences of farmers in Europe with the process of ensiling and also of the results of his own experiments. The following is quoted from this letter: Last season at Champaign, 111., experiments on a small scale were made under my direction in the ensilage of corn stalks and broom corn seed with results that were on the whole satisfactory. Two pits 12 feet long, 6 feet wide, and about 8 inches below the surface of the soil were filled with cut corn stalks of a late variety and the piles carried up as high as the stalks could readily be made to keep their place. A covering of straw about 4 inches thick was then put over the pile and about 6 to 8 inches of earth added. After two or three weeks when the pile had settled an additional layer of earth about 8 inches in thickness was added. The stalks were cut by hand with a very inferior straw cutter. Professor Miles went on to state that one of the pits of corn stalks was opened in December and that there was a layer of rotted material about 3 inches thick. The second pit was opened March 13, 1876, and considerably more had decayed. Below this decayed layer the silage had kept perfectly. Similar results were obtained in the case of the broom corn seed. Professor Miles suggested that the term ensilage be adopted to designate the method as there was no English equivalent. Francis Morris, of Oakland Manor, near Ellicott City, Maryland, built a silo in 1876 and filled it with corn. This silo was a trench 4 feet wide, 10 feet deep, and 24 feet long. Mr. Morris became a very carrier: history of the silo. i 8 i enthusiastic advocate of the silo and his experiences were given in a number of farm papers. Mr. J. B. Brown, of New York, translated the book " Ensilage of Maize " of A. Goffart. This was published in 1879 an d distributed largely as an advertisement of an implement company of which Mr. Brown was president. This little book created a great deal of interest and Goffart was heralded as the " Father of Ensilage. " From that time until the present, books, bulletins, and articles in the agricultural press on the subject "silos and silage" — much of it controversial — have appeared so frequently that the litera- ture is too voluminous to need special enumeration. For the first 20 years after its introduction, silo-building was a haphazard proposition. Each farmer tried his own individual ideas. Some of these were good, others were costly experiences to the builders. On the whole many valuable data were gained during this experimental stage. It was soon found that with air-tight walls all of the spoiled silage was at the top. This led to building the silos above ground and taller. With the taller silos weighting on the top of the silage was unneces- sary. There was also a big saving in labor of tramping the cut ma- terial at filling time. In 1 891, Prof. F. H. King, of the Wisconsin Agricultural Experi- ment Station, began a study of the whole subject of silo construction and ensiling. This study covered a number of years in which he personally inspected over 200 silos. He was aided by quite a mass of data which had been collected at other experiment stations. The publication of Professor King's researches marked a new era in silo construction (7, 8). A few cylindrical silos had been constructed be- fore that time, but King's description of the details of construction were so clear that the Wisconsin silo became for a decade the most popular type built. The pits, trenches, and low, squatty, rectangular structures gave way to the tall, cylindrical silos. The Wisconsin silo is no longer practicable owing to the comparative high cost of con- struction and early decay. But the principles of construction, the weights and lateral pressure of silage at different depths, and the size of silo to build for a given number of animals as worked out by King and published in several reports and bulletins of the Wisconsin Agri- cultural Experiment Station are used more today than when they were first written. King's silage tables are classics. No man has done more than he to make the silo a success. Literature Cited. 1. C , G. Sour-fodder making. In Amer. Agr., v. 32, no. 10, p. 370. 1873. 2. Fouvielle, . Presse Scientific, t. 1, p. 162. Paris, 1865. 1 82 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. 3. Goffart, A. The Ensilage of Maize. Trans, by J. B. Brown. New York, 1879. 4. (Harrison, Fairfax.) Roman Farm Management, by " A Virginia Farmer," p. 168. New York, 1913. 5. Jenkins, H. M. Report on the practice of ensilage, at home and abroad. In Jour. Royal Agr. Soc. (England), 2d ser., v. 20, p. 130. 1884. 6. Johnston, J. F. W. On the feeding qualities of the natural and artificial grasses in different states of dryness. In Trans. Highland and Agr. Soc, 1843-1845, p. 57. 1846. 7. King, F. H. The construction of silos. Wis. Agr. Expt. Sta. Bui. 28. 1891. 8. . The construction and filling of a round silo, 16 feet outside diam- eter and 27 feet high. Capacity 80 ton. In 9th Ann. Rpt. Wis. Agr. Expt. Sta. 1891/92, p. 121-128. 1893. 9. Miles, Manley. Silos, Ensilage and Silage, p. 34. New York, 1889. 10. Morris, F. Preserving green corn fodder.. In Amer. Farmer, v. 6, no. 3, p. 94. 1877. 11. Young, Arthur. Annals of Agriculture, v. 1, p. 207. London, 1786. Also the following unsigned articles : 12. Curing green fodder. In Amer. Agr., v. 34, no. 6, p. 222. 1875. 13. . In Amer. Farmer, v. 4, no. 6, p. 250. 1875. 14. Purchas his Pilgrimes, a Relation of Plymouth, v. 19, p. 320. Glasgow, 1889. 15. French mode of curing forage. In Rpt. U. S. Com. Agr. for 1875, p. 387- 408. Washington, 1876. AGRONOMIC AFFAIRS. 183 AGRONOMIC AFFAIRS. MEMBERSHIP CHANGES. The membership reported in the April Journal was 531. Since then, seven new members have been added, making the present mem- bership 538. The names and addresses of the new members, with the changes of address which have been noted by the secretary and editor, are as follows: New Members. Anderson, Harold, 610 Keyser Bldg., Baltimore, Md. Christensen, Fred G., Box 785, Kingsbury, Cal. De Young, William, Soils Dept., Univ. of Mo., Columbia, Mo. Ogaard, A. J., College of Agriculture, Bozeman, Mont. Qua, N. C, School of Agriculture, Vermilion, Alta., Canada. Singleton, H. P., Washington State College, Pullman, Wash. Sumner, Herbert R., College of Agriculture, Bozeman, Mont. Changes of Address. Bancroft, Ross L., Phillips, Wis. Cunningham, C. C, R. F. D. No. 4, Edwards, Kans. Schuster, Geo. L., Experiment Station, Newark, Del. NOTES AND NEWS. L. R. Breithaupt, formerly superintendent of the Harney County Branch Station, Burns, Oreg., and more recently engaged in farming in Idaho, is now county agricultural agent of Malheur County, Oreg. T. S. Buie, assistant agronomist at the Georgia station, has been made head of the department of agronomy. W. L. Burlison, for the past several years professor of crop pro- duction in the University of Illinois, has been made head of the de- partment of agronomy, filling the vacancy caused by the death of Dr. C. G. Hopkins. J. B. R. Dickey, formerly extension specialist in agronomy and soils in Rutgers College, is now assistant professor of agronomy ex- tension in the Pennsylvania college. J. A. Evans, for the past several years connected with the Federal Office of Extension Work in the South, has been appointed to suc- ceed Bradford Knapp at the head of that office. I84 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Sidney B. Haskell, head of the department of agronomy at the Massachusetts college and station from 191 1 to 1916 and since con- nected with the Soil Improvement Committee of the National Ferti- lizer Association, has been elected director of the Massachusetts station, effective July 1. R. R. Hudelson, assistant professor of soils, and E. M. McDonald, assistant professor of farm crops in the University of Missouri, have resigned to engage in farming in Alberta. James T. Jardine, for the past several years in charge of grazing experiments in the Forest Service, has been elected director of the Oregon station, effective June I. Harry L. Kent, director of agricultural education in Kansas under the Smith-Hughes act and principal of the school of agriculture at the Kansas College, has been elected head of the Fort Hays Branch Station, Hays, Kans., effective May 15. J. D. Luckett, formerly specialist in field crops on the Experiment Station Record, is now editor and librarian of the New York State station. George L. Schuster, formerly assistant agronomist at the Uni- versity of West Virginia, has been agronomist of the Delaware col- lege and station since April I. John A. Slipher, formerly of the department of soil technology at Purdue University, is now assistant manager of the agricultural de- partment of the National Lime Association. S. H. Starr, formerly assistant professor of farm management at the Georgia college, is now director of the newly established Georgia Coastal Plains station at Tifton, which is entirely distinct from both the Georgia college and station. E. D. Stewart, superintendent of the Langdon (N. Dak.) experi- ment farm and recently appointed superintendent of farm demon- strations in North Dakota, died of influenza March 12. Theodore Stoa, formerly scientific assistant in the Federal office of cereal investigations, is now assistant agronomist at the North Dakota station. John H. Voorhees, assistant agronomist in extension at Cornell University, resigned February 1 to become associate editor of the Pennsylvania Farmer. JOURNAL OF THE American Society of Agronomy Vol. 12. September-October. 1920. Nos. 6-7 THE INEQUALITY OF RECIPROCAL CORN CROSSES. 1 Frederick D. Richey. Introduction. Although reciprocal crosses are in general practically indistinguish- able, there are well recognized exceptions to this rule. For example, reciprocal crosses between Digitalis purpurea and D. lutea constantly produce hybrids that resemble the female parent, whereas reciprocal crosses between Oenothera biennis and O. muricata give hybrids that strongly resemble the male parent. Reciprocal crosses between varie- ties or strains of corn (Zea mays L.) have been compared by several investigators and different results obtained. Shull (13) 2 concluded from the results of his investigations that u The reciprocals between two distinct, self-fertilized families are equal," and East and Hayes (4) note the agreement of their results with those of Shull. Burtt-Davy (2) states with reference to the crosses between breeds with 8 and 18 rows, "The ears produced by the cross and the reciprocal cross are indistinguishable/' and (3) observes in general, "It appears to.be immaterial which breed fur- nishes the male and which the female parent ; the results in the F 2 generation are usually the same in either case." Williams and Wel- ton (15) note without comment the equality of two reciprocal crosses, both in yield and in moisture content. As to the inheritance of indi- vidual qualitative factors, it seems sufficient to say that no report has 1 Contribution from the Office of Cereal Investigations, Bureau of Plant Industry, United States Department of Agriculture, Washington, D. C. Pub- lication approved by the Secretary of Agriculture. Received for publication April 22, 1920. 2 Reference is by number to " Literature cited," p. 195. 1 86 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. been found of inequality in germinal inheritance between reciprocal crosses. On the other hand, however, McCluer (12) noted that the cross, Queen Golden X White Dent, resembled the dent parent, whereas the reciprocal cross was intermediate. Gernert (5) states that differ- ences between reciprocals were observed both as to plant and seed characters, and more specifically notes differences of four to eight days in the period from planting to silking. Jones (7) concludes, " Although reciprocal crosses are on the whole nearly equal in respect to the degree in which heterosis is shown, there is some evidence from Table 12 that this is not always so." Jones' conclusion is of particu- lar interest, as he used later generations of some of the inbred strains used by East and Hayes (4) in their investigations. Jones notes a correlation between better seed development in certain strains and larger yields from the crosses having these strains for pistillate parents, and is inclined to attribute the differences between the re- ciprocals to the difference in the food materials furnished the young plants. He notes, however, " It is not certain that these differences can be accounted for on a purely nutritional basis," and suggests the " possibility of unequal germinal reactions with different cytoplasms." Jones (7, Table 17, p. 51) also gives data which show inequality be- tween reciprocals in each of two pairs of reciprocal crosses between different types of corn. The present paper compares the yields of three pairs of reciprocal crosses between commercial varieties of corn, and shows inequality in each pair. Data also are presented which show that in one pair of these reciprocals the inequality is caused by the inheritance -of some character from one of the parent varieties only when that variety is used as the staminate parent. Comparison of Reciprocals. 1 ' . » VARIETIES AND METHODS. The following varieties of corn were used as parents of the recip- rocal crosses discussed : U. S. Selection No. 119, a strain of Boone County White, grown near Washing- ton, D. C, since 1902. U. S. Selection No. 120, a cross of Hickory King X No. 119, grown near Wash- ington, D. C, since 1902. U. S. Selection No. 199, Fraley Yellow Dent, grown near Derwood, Md., since 1900. U. S. Selection 200, St. Charles White, grown near St. Charles, Mo., since 1850. "Johnson," a mixed strain of Johnson County White, grown in northeastern Arkansas since 1912. RICHEV : RECIPROCAL CORN CROSSES. 187 The crosses were made by growing rows of the varieties used as pistillate parents between rows of the varieties used as staminate parents, and detasselling the former rows before they had shed pollen. From 15 to 25 ears were obtained from the detasseled rows to repre- sent each cross. The methods of comparison are given in connection with each ex- periment. In all the experiments, however, the corn was planted thick and was later thinned to the desired stand. All yields were cor- rected to a basis of air-dry shelled corn. RECIPROCAL CROSSES BETWEEN U. S. SELECTION NOS. 120 AND I99. These crosses were compared at Gaithersburg, Md., and at Occo- quan, Va., in 1912. 3 Each cross was grown in a row between rows of the parent varieties, and intra-hill with each parent variety. The reciprocals therefore can be compared thru the ratios of their re- spective yields to the average yield of the parent varieties. This com- parison is made in Table 1, the yield per plant being given for reference. Table i. — Comparison of the reciprocal crosses between U. S. Selection A?os. 799 and I20. a Locality and method of planting. Replication No. 120 X 199. 199 x 120. 120 x 199* exceeds its reciprocal.. Yield per plant. Ratio of cross to parents, Yield per plant. Ratio of cross to parents. Average of parents. Cross. Average of parents. Cross. Pounds. Pounds. Percent. Pounds. Pounds. Percent. Per cent. b Gaithersburg: Inter-row . . • I 0.631 0.681 107.9 0.462 0.461 99-8 8.1 2 .615 .692 112. 5 •532 •530 99.8 12.7 Ave. .623 .687 110.3 •497 .496 99-8 10.5 Intra-hill . . 1 .613 .676 110.3 •494 •443 89*-7 23.0 2 .664 .698 105. 1 •538 .519 96.5 8.9 Ave. •639 .687 107.5 .516 .481 93-2 15.3 Occoquan: Inter-row. . 1 •595 .567 95-5 •578 .527 91.2 4-5 2 •473 •45i 95-3 .507 •517 102.0 -6.6 Ave. •534 •509 95-4 • 543 .522 96.1 -0.7 Intra-hill . . 1 .566 •579 102.3 •493 .522 105.9 -3-4 2 •435 .440 IOI.I .487 •542 111.3 -9.2 Ave. .501 .510 101.8 .490 •532 108.6 -6.3 u All averages are calculated directly from basic yield figures. 6 Percent of 199 X 120, in terms of the average of the parents. 3 These crosses were grown in connection with experiments to determine the- relative productiveness of Fj crosses and their parent varieties. These experi- ments were planned and carried out in 1911-1912 by the Office of Corn Investi- gations, U. S. Department of Agriculture. The methods used in these experi- ments are given more fully in a later part of this paper. 1 88 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. The results show that No. 120 X No. 199 was markedly superior to its reciprocal at Gaithersburg, whereas at Occoquan it was slightly inferior to its reciprocal. EXPERIMENTS IN I915-I917. The experiments from 191 5 to 1917 were conducted at Armorel, Ark. Crosses were made between U. S. Selection Nos. 119 and 120 in 191 5. Two sets of ears were used to represent each variety, and each set was used as a staminate and as a pistillate parent. Sixteen com- binations were thus obtained ; four to each parent and cross. Each of these 16 combinations was grown in 19 16 in a 2-row plat between i-row check plats. The actual yields of the experimental plats were corrected by the formula, (a + b)/(c' + c") X C, in which a and b are the yields of the two rows of an experimental plat, c' and c" are the yields of the check plats on each side of a and b, and C is the average of all the checks. The four combinations that represented each variety or cross are treated as replications. The same methods were used to make and compare the crosses between Johnson and U. S. Selection No. 200 in 1916-17. The results of these experi- ments are shown in Table 2. Table 2. — Comparison of reciprocal crosses between U. S. Selection Nos. 119 and 120, and betzveen Johnson and U. S. Selection No. 200, in bushels of shelled corn per acre. NO. 119 X NO. 120. Vield in bushels per acre. Strain. Replication No. No. of Average < plants. 1 2 3 4 35-o 34-6 31-4 31.0 33-o±o.7 1,087 120 X 119 40.0 39-9 39-2 38.1 39-3 ± -3 1,107 35-9 35.5 34-5 33.7 34-9 ± -3 1,104 No. 119 32.9 27.1 31-3 28.5 30.0 ± .9 I,Il6 JOHNSON X NO. 200. Johnson .... 61.5 60.9 56.7 59-1 59.6 ±0.7 1.392 Johnson X 200 69.1 66.8 67.4 68.1 67-9 ± -3 i,392 200 X Johnson 64.6 67.8 63.8 65.7 65. 5± -6 1-392 66.2 63.0 61.7 61.5 63.1 ± -7 1.392 The cross No. 120 X No. 119 exceeds its reciprocal by 4.4 ± 04 bushels per acre, or 12.6 ± 1.2 per cent. The cross Johnson X No. 200 exceeds its reciprocal by 2.4 ± 0.7 bushels per acre, or 3.7 ±1.0 per cent. The difference between the reciprocals is significant in each RICHEY : RECIPROCAL CORN CROSSES. 189 case, that between No. 120 X No. 119 and No. 119 X No. 120 being especially so. In this connection it is of interest to note that No. 120 X No. 119 is the same combination that gave the maximum in- crease over the better parent in earlier crossing experiments (6). Cause of Inequality Between Reciprocals. Nutritional differences and varying germinal reactions with differ- ent cytoplasms have been suggested as possible causes of inequality between reciprocal corn crosses. Relative early vigor and productive- ness of the parents is the only evidence that bears even indirectly on these points in connection with the present crosses. The use as a specific parent of the variety producing either the more vigorous early growth or the larger yield had no consistent effect in determining the better reciprocal. This is shown in Table 3. Table 3. — Relation between the better reciprocal and greater early vigor or productiveness in a specific parent. Parent showing more vigor- Parent producing Better reciprocal. ous early growth. greater yield. No. 120 X No. 119 '. . . . Pistillate Pistillate Johnson X No. 200 Staminate Staminate No. 120 X No. igg a Neither Staminate a n At Gaithersburg. The possibility of sex linkage as a cause of inequality between the reciprocals of Nos. 120 and 199 is interesting in the light of the fol- lowing experiments. Two series of crosses were made in 191 i, 4 No. 120 being used as the staminate parent of one series at Vienna, Va., and No. 199 of the other at Derwood, Md. Twenty varieties, includ- ing Nos. 120 and 199, were used as pistillate parents and seed of the following classes thus obtained : Nos. 120 and 199 from detasseled plants; the reciprocal crosses between Nos. 120 and 199; and 18 varieties X No. 120 and X No. 199. These, together with the pistil- late parent varieties, were compared at Gaithersburg, Md., and Occo- quan, Va., in 1912. The arrangement was the same at both places. Each cross was grown with its parents in a group of six rows in the following order : 1. Staminate parent. 2. Cross. 3. Pistillate parent. 4. Cross and staminate parent. 5. Cross and pistillate parent. 6. Pistillate parent and staminate parent. 4 See footnote 3. I9O JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. In the first three of these rows comparisons are based on row yields and are referred to as " inter-row." In the last three rows the two classes that were grown in different sides of the hills of the same row are compared and these comparisons are referred to as " intra-hill " (10, 11). Each row contained 60 2-plant hills and was duplicated. For convenience, the 18 varieties X 120, the 18 varieties X 199, and the 18 pistillate parent varieties will be referred to as the 120-crosses, the 199-crosses, and the 18 varieties, respectively. The 120-crosses were compared with their respective parents in a plat of 18 consecutive groups. No. 120 occupied row 1 of each of these groups, one of the 120 crosses occupied row 2 of each group, and one of the 18 varieties occupied row 3 of each group. The rela- tive average yield of No. 120 and of the 120-crosses therefore can be computed as percentages of the average yield of the 18 varieties as illustrated in the formula, 1 1 —J— j 2 — U 1 3 — I U I 18 — - : — r^ = Relative average yield of No. 120, 3 1 + 3 2 + 3 s -\ h3 18 y in which the terms in the numerator represent the yields of rows of No. 120 in successive groups, and those in the denominator represent Table 4. — Comparison of iog-crosses and 120-crosses thru their ratios to 18 varieties. 0. Locality and method of planting. Replication No. 199-crosses. 120-crosses. 199-crosses exceed 120- crosses. Yield per plant. Ratio of crosses to varieties. Yield per plant. Ratio of crosses to varieties Varieties. Crosses. Varieties. Crosses. Pounds. Pounds. Percent. Pounds. Pounds. Percent. Percent}* Gaithersburg: Inter-row . . I 0.511 0.592 H5-9 0.466 0.499 107. 1 8.2 2 .536 .618 H5-3 •521 ■543 104.2 10.7 Ave. • 524 .606 115. 6 •494 .521 105-5 9.6 Intra-hill . . 1 •5ii •579 "33 .487 .486 99-8 13-5 2 •534 • 593 III.O •544 .527 96.9 14.6 Ave. • 523 .586 112. .516 •507 98.3 13-9 Occoquan: Inter- row . . 1 .467 •495 106.0 •495 .512 103.4 2-5 2 •452 .476 105-3 .538 •544 IOI.I 4.2 Ave. .460 •485 105.4 •517 .528 102. 1 3-2 Intra-hill . . 1 •483 •513 106.2 •5ii • 532 104.1 2.0 2 •452 .480 106.2 •534 .548 102.6 3-5 Ave. .467 .496 106.2 •523 •540 103.2 2.9 All averages are calculated directly from basic yield figures. 6 Percent of the 18 varieties X No. 120, in terms of the 18 varieties. the yields of the 18 varieties in the same groups. The 199-crosses were grown in a similar plat of 18 groups and the relative average RICHEY : RECIPROCAL CORN CROSSES. yield of No. 199 and the 199-crosses can also be computed as per- centages of the average yield of the 18 varieties. Under the inter- row method, No. 120 and the 120-crosses then can be compared with No. 199 and the 199-crosses thru their relative yields. Similarly under the intra-hill method, Nos. 120 and 199 can be compared thru the relation of their yields to the yield of 18 varieties grown intra-hill with them in row 6 of each group ; and the 120-crosses can be compared with the 199-crosses thru their relative yields in row 5 of each group. Table 4 shows the comparison of the crosses and Table 5 the comparison of Nos. 120 and 199. The actual yields in pounds of air-dry shelled grain per plant are shown for reference. At Gaithersburg No. 199 exceeded No. 120 by 14.7 percent, when grown inter-row, and by 28.7 percent, when grown intra-hill. At Occoquan the differences were negligible, being but 2.1 percent and 0.6 percent. Table 5. — Comparison of Nos. 199 and 120, thru their ratios to 18 varieties.® Locality and method of planting. Replication No. No. 199. No. 120. No. 199 exceeds No. 120. Yield per plant. Ratio of No. 199 to 18 varieties. Yield per plant. Ratio of No. 120 to 18 varieties. Average of 18 varities. No. 199. Average of 18 varieties. No. 120. Pounds. Pounds. Percent. Pounds. Pounds. Percent. Percent} Gaithersburg: Inter-row . . I 0.511 0.611 119. 6 0.466 0.492 105.6 13-3 2 .536 .628 117. 2 .521 •525 100.8 16.3 Ave. •524 .619 118. 1 •494 •509 103.0 14.7 Intra-hill . . 1 .487 .630 129.4 .501 .490 97.8 32-3 2 •53i .650 122.4 •539 .526 97.6 25-4 Ave. •509 .640 125-7 • 520 .508 97-7 28.7 Occoquan: Inter- row . . 1 .467 •504 107.9 •495 .522 105-5 2.3 2 •452 .469 103.8 .538 •551 102.4 1.4 Ave. .460 .487 105.9 •517 .536 103-7 2.1 Intra-hill . 1 •493 .487 98.8 .508 • 512 100.8 — 2.0 2 •448 •459 102.5 •542 •537 99-3 3-2 Ave. .471 •473 100.4 .525 .524 99.8 0.6 a All averages are calculated directly from the basic yield figures. 6 Percentage of No. 120 in terms of the 18 varieties. At Gaithersburg the 199-crosses exceeded the 120-crosses by 9.6 percent, when grown inter-row, and by 13.9 percent, when grown intra-hill. At Occoquan, on the other hand, the 199-crosses exceeded the 120-crosses by only 3.2 percent, inter-row, and 2.9 percent, intra- hill. This method of comparison thru relative yields differs from the 192 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. ordinary method of checking for environmental variation only in the use of 18 varieties as the standard instead of one variety. The use of a number of varieties seems preferable, for the average yield of a number of varieties will be less influenced by specific adaptation and therefore will represent better the potential productiveness of the environment. Nos. 120 and 199 also were compared in six pairs of adjacent rows, and intra-hill in six rows at Gaithersburg and at Occoquan. The direct comparison between the yields of No. 120 and No. 199 in these rows is made in Table 6. This agrees in general so closely with the comparison thru relative yields (Table 5) as to demonstrate the reliability of the latter method. Table 6. — Comparison of Nos. 120 and 199, based on actual yields' in pounds of air-dry shelled corn per plant in adjacent rows and intra-hill. GAITHERSBURG. Comparison Inter-row. Intra-hill. No. No. 199. No. lao. No. 199 exeeeds No. 199. No. 120. No. 199 exceeds No. 120. No. 120. I O.467 0-457 O.OIO O.732 0-594 O.I38 2 .568 •494 •074 .724 .620 .104 3 .681 .580 .101 •570 .418 .152 4 .671 •558 • 113 .602 •458 .144 5 .646 .561 .085 .714 •530 .184 6 •654 •572 .082 .680 •574 .106 Average .... O.615 0-537 0.078 ±.010 O.670 0.532 O.136 ±.008 OCCOQUAN. I 0.575 0.581 —0.006 0.592 0.492 0.100 2 .490 •524 - .034 •392 •430 - .038 3 .643 . -547 .096 •544 .526 .018 4- •473 •472 .001 .528 •534 — .006 5 .566 •594 - .028 .648 •534 .114 6 • 590 .561 .029 .662 .598 .064 Average .... 0.556 0-547 o.oio±.oi3 0.561 0.51.9 0.042 ±.017 No. 199 exceeded No. 120 by 0.078 ±0.10 bushel, or 14.5 ± 1.9 percent, at Gaithersburg when grown inter-row, and by 0.136 ± .008 bushel, or 25.6 ± 1.5 percent, when grown intra-hill. At Occoquan, it exceeded No. 120 by .010 ± .013 bushel, or 1.8 ± 2.4 percent, when grown inter-row, and by 0.042 ± .017 bushel, or 8.1 ± 3.3 percent, when grown intra-hill. The effect of greater competition in the intra-hill method reported by Kiesselbach (10) as a source of error in varietal experiments is richey: reciprocal corn crosses. 193 clearly shown in these results. This effect serves to strengthen the conclusiveness of the present results, however, as the intensification is entirely consistent with the results obtained under both methods. These experiments show conclusively that No. 199 was superior to No. 120 (Tables 5 and 6) and that the 199-crosses were superior to the 120-crosses (Table 4) at Gaithersburg. At Occoquan the differ- ences were insignificant. That No. 120 X No. 199 was better than OCCOQUAN Intra- Inter- hill row GAITHERSBURG Inter- Intra- row hill 130 125 120 115 lip 105 100 95 90 199 120 120X199 199X120 . 199-CRQSSES 120-CROSSES Fig. 14. Graph showing the ratios of No. 199 to No. 120, No. 120 X No. 199 to No. 199 X No. 120, and the 199-crosses to the 120-crosses, at Occoquan and Gaithersburg. No. 199 X No. 120 at Gaithersburg and not so good at Occoquan has been shown in Table 1. This similarity in the relative behavior of the series that had No. 199 for staminate parent in comparison with the series, that had No. 120 for staminate parent is brought out more clearly in Table 7, which summarizes Tables I, 4, 5, and 6, and in figure 14, which shows the relation graphically. These two series differed thruout in only one way other than parentage, namely, the conditions under which the seed was grown. 194 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Table 7. — Superiority of No. 120 X No. 199, the 199-crosses, and No. 199, over No. 199 X No. 120, the 120-crosses, and No. 120, respectively. (Summary of Tables 1, 4, 5, and 6.) Superiority of : Table No. Gaithersburg. Occoquan. Inter-row. Intra-hill. Inter-row. Intra-hill. Percent. Percent. Percent. Percent. 120 X 199 I 10.5 15-3 -0.7 -6.3 199-crosses 4 9.6 13-9 3-2 2.9 199 (relative) 5 14.7 28.7 2.1 0.6 199 (actual) 6 14.5 ±1.9 25.6±i.5 1.8 ±2.4 8.1 ±3.3 The seed of No. 120 and the crosses of which it was the staminate parent was grown at Vienna, Va., and seed of No. 199 and the crosses of which it was the staminate parent at Derwood, Md. These places are about 20 miles apart, the soils and seasons are similar in general, and there was no apparent difference in the quality of the seeds as is seen from Table 8. Table 8. — Notes on seed grown at Vienna, Va., and Derwood, Md., in ign, based on the average of 19 crosses and the staminate parent. Locality. Staminate parent. Seed condi- tion.** Germination. Weight of struck 2 quarts Percent. Percent. Pounds. No. 120 79.8 98.95 3-32 No. 199 81.5 98.65 3.28 a Seed condition represents the judgment of the observer. From the data in Table 8 there is nothing to indicate that the dif- ferences in yield were caused by the difference in the conditions under which the seed was grown. Moreover, under both the inter-row and the intra-hill methods and in the difference between the methods, the 199-crosses show this suitability to Gaithersburg conditions less than does No. 199. This indicates an intermediate condition and genetic causes rather than physiological ones. The only logical interpretation of the above facts and one which follows them so closely that it is essentially equivalent to restating them is ; that No. 199 possessed some character that particularly suited it to the Gaithersburg environment. No. 120 did not possess this character, nor did the 18 varieties as a whole. This character was transmitted, in part at least, to the 199-crosses and to No. 120 X No. 199, in each of which No. 199 was the staminate parent. This char- acter was transmitted slightly, if at all, to No. 199 X No. 120, in which No. 199 was the pistillate parent. There was, therefore, a dif- RICHEY : RECIPROCAL CORN CROSSES. 195 ference in transmission from No. 199 when it was used as the stami- nate and the pistillate parent, respectively. Unfortunately, this is as far as the facts go and any attempt to explain the exact method of this unequal transmission must be specu- lative. However, the results follow the operation of sex-linked in- heritance so closely that we are led to consider the possibility of some type of sex linkage as a cause of inequality between reciprocal corn crosses. That sex-linked inheritance of specific qualitative characters has not been observed in corn is relatively unimportant ; for nine of the factors more thoroly investigated already have been placed in but three groups (8, 9), whereas the number of chromosomes in corn is large. That sex linkage may occur in corn does not seem impossible in view of its occurrence in Lychnis (13) and in Begonia (1). Conclusion. The reciprocal crosses between varieties or strains of corn some- times are unequal. A difference in the food materials furnished the young plants by the different maternal parents and a difference in germinal reactions with different cytoplasms have been suggested as possible causes of such inequalities. That some type of sex-linked inheritance must at least be considered as a possible cause of in- equality between rciprocal corn crosses is shown by the unequal trans- mission from No. 199 as staminate and as pistillate parent in the above experiments. Literature Cited. 1. Bateson, W., and Sutton, Ida. Double flowers and sex linkage in Begonia. In Jour. Genetics, 8: 199-207. 1919. 2. Burtt-Davy, J. Observations on the inheritance of characters in Zea mays Linn. In Trans. Roy. Soc. South Africa, v. 2, pt. 3, p. 261-270. 1912. 3- . Maize, Its History, Cultivation, Handling and Uses. London, New York, etc. 1914. 4. East, E. M., and Hayes, H. K. Heterozygosis in evolution and in plant breeding. U. S. Dept. Agr., Bur. Plant Indus. Bui. 243. 1912. 5. Gernert, W. B. The analysis of characters in corn and their behaviour in transmission. Champaign, 111. 1912. 6. Hartley, C. P., Brown, E. B., Kyle, C. H., and Zook, L. L. Crossbreeding corn. U. S. Dept. Agr., Bur. Plant Indus. Bui. 218. 1912. 7. Jones, D. F. The effects of inbreeding and crossbreeding upon develop- ment. Conn. Agr. Exp. Sta. Bui. 207. 1918. 8. , and Gallastegui, G. A. Some factor relations in maize with refer- ence to linkage. In Amer. Nat., 53 : 239-246. 1919. 9. Kempton, J. H. Inheritance of spotted aleurone color in hybrids of Chi- nese maize. In Genetics, 4: 261-274. 1919. I96 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. 10. Kiesselbach, T. A. Studies concerning the elimination of experimental error in comparative crop tests. Nebr. Agr. Expt. Sta. Research Bui. 13. 1918. 11. Kyle, C. H. Directions to cooperative corn breeders. U. S. Dept. Agr., Bur. Plant Indus. Pub. 564. 1910. 12. McCluer, G. W. Corn crossing. 111. Agr. Expt. . Sta. Bui. 21. 1892. 13. Shull, G. H. Hybridization methods in corn breeding. In Proc. Amer. Breeders' Assoc., 6: 63-72. 1910. 14. . Sex-limited inheritance in Lychnis dioica. In Zeits. f. Induk. Abs. Verer., 12: 265-302. 1914. 15. Williams, C. G.,. and Welton, F. A. Corn experiments. Ohio Agr. Expt. Sta. Bui. 282. 1915. RELATIVE YIELDS FROM BROKEN AND ENTIRE KERNELS OF SEED CORN. 1 Ernest B. Brown. It is not uncommon to find kernels in seed corn that have been physically injured. These injuries are usually the result of weevil at- tacks, gnawing by mice, cracking in shelling and seeding, or abrasions from other causes. Ordinarily, the presence of a small percentage of injured kernels is not regarded as seriously affecting the quality of the seed. Occasionally, either from necessity or otherwise, corn con- taining a considerable percentage of injured kernels is planted. Whether injuries of this physical nature affect the growth, develop- ment, and productivity of the plant is the subject of this investigation. Three ears of U. S. Selection No. 120, a white dent variety, were used in the experiment. Broken or cracked kernels from each ear were planted in comparison with uninjured kernels from the same ear. The breaking was similar to the injuries that frequently occur in mechanical shelling and seeding, and analogous to weevil and mouse damage in that the amount of nutriment in the kernel was lessened. The injuries were confined to the endosperm, in no instance extending to the germ. The experiment was conducted on Arlington Farm, near Washing- ton, D. C, in 1914. The corn was planted May 29 by hand in rows 3.3 feet apart, 40 hills per row, with the hills 3.3 feet apart in the row. When the plants were 6 to 8 inches tall, the stand was thinned to 2 plants per hill. The crop was harvested October 30. The statis- tical data are presented in Table 1. 1 Contribution from the Office of Cereal Investigations, Bureau of Plant Industry, United States Department of Agriculture, Washington, D. C. Re- ceived for publication June 30, 1920. brown: yields from entire kernels of seed corn. 197 Table i. — Comparison of yields from broken and entire kernels of seed corn, U. S. Selection No. 120. Row Nq. Class. Adult plants. Ears pro- duced. Total weight of ears. Ears per plant. Average weight of ears. Yield per plant. Yield per acre". Pounds. Pounds. Pounds. Bushels. Ear 1: A. A 04 Broken seed 43 52 29.1 1 .2 10 O.560 0.677 67.7 Kntire seed 77 /6 r-A A 50.0 A oo .900 •745 •735 73-5 Ear 2: 66 Broken seed 66 57 38.8 .864 .681 .588 58.8 67 Entire seed 86 82 59.8 •954 • 730 .696 69.6 Ear 3: 68 Broken seed 62 62 41.6 1. 000 .671 .671 67.1 69 Entire seed 69 71 49.8 .1.029 .702 .722 72.2 Averages Broken seed 57 57 36.5 1. 000 .640 .640 64.0 Entire seed 78 77 55-4 .988 .716 .716 71.6 Difference 7.6 a Based on 7,000 plants per acre and 70 pounds of ears to the bushel. The field germination from the broken seed was less than from the entire seed and the seedlings were weaker. The general height of plants at maturity did not materially differ. Apparently the mutilation of the seed was not a limiting factor in the height of the plant. In number of ears per plant the broken seed exceeded the entire seed in one out of three comparisons and in the general average. The increase in number of ears was associated with the increased space per plant resulting from the poorer stands from broken seed. In weight of ear and yield per plant, the yield from broken seed was consistently less than that from the entire seed. On an acre-yield basis, the broken seed yielded 7.6 bushels less than the entire seed. In practical corn growing, the losses in yield prob- ably would be much less, frequently negligible, but the tendency would be the same. I98 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. AN OUTLINE OF AN UNDERGRADUATE COURSE IN GRAIN GRADING. 1 John B. Wentz. One phase of agronomic work which has come to the attention of teachers of farm crops in the last few years is the teaching of the grading of grains according to the Federal standards established by the Grain Standards Act. As the grade requirements of the different grains have been published by the Bureau of Markets, it has been found necessary to change the system of teaching the grading of grains to conform to the requirements of the new standards. In in- stitutions where no special attention had previously been given to market grading of grains, it has been found necessary to give con- sideration to this work. The use of the Federal standards in the grain markets has stimulated a desire among grain dealers and farmers to gain a knowledge of the methods of applying the standards, and this fact, together with the generally increased interest in the last few years in marketing of farm products, has greatly increased the demands made on the agricultural colleges for information on market grading of grains. Numerous applications were addressed to the Bureau of Markets by farm crops teachers asking for information and materials to be used in instructional work. These requests finally resulted in a conference -of the agricultural college men interested in grain grading and repre- sentatives of the Bureau of Markets, to furnish subject matter on the Federal standards for instructional work. This conference was held .at the General Field Headquarters of Federal Grain Supervision, Chicago, 111., September 8 to 11, 1919. In the two years previous to the conference' at Chicago, the agronomy department of the Maryland State College had been col- lecting the apparatus and subject matter for the teaching of grain grading according to the Federal standards, and in the year previous to the conference a 2-hour course was offered to juniors in the winter term. This year the course is being offered in the same term, but the number of hours has been increased to three, including two lectures and one 3-hour laboratory period per week. The interest of the stu- 1 Contribution from the Department of Agronomy, Maryland State College, College Park, Maryland. Received for publication April 12, 1920. WENTZ : COURSE IN GRAIN GRADING. 199 dents in this course and the amount of general interest it has con- tributed to the farm crops work at this institution seems to justify a report to members of other agronomy departments who might be in- terested in this work. Two views of the apparatus used are shown in Plate 8. Following is an outline of the course as it is now being conducted. Outline of Course in Grain Grading. I. History leading up to the passing of the Grain Standards Act ( 7 ). 2 II. United States Grain Standards Act (2). A. Date passed. B. Value in foreign trade. C. Value in interstate trade. D. Indirect effect upon intrastate and local trade (6). E. How the farmer is affected (1). F. Classifications and grades provided by the Federal standards. 1. For wheat (3). a. Classes, subclasses, and grades. (As shown by charts put out by the Bureau of Markets.) b. Wheat districts (shown by map). c. Descriptions of the six market classes of wheat (both thrashed and head samples of each class exhibited. District where grown. Texture. Color. Milling quality. 2. For corn (3). a. Classes and grades (as shown by charts). 3. For oats (4). a. Classes and grades (as shown by charts). G. The organization to carry out the provisions of the act. 1. The licensed inspector. a. Qualifications of the inspector. b. Duties. c. How he obtains his license and his relation to the Federal government (2). 2 Reference by number is to " Literature cited," p. 203. 200 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. d. Rules and regulations governing the inspec- tor (2). 2. The district supervisor. a. Division of United States into supervision dis- tricts (shown by map). b. Qualifications of the supervisor. c. Duties. d. How to use the supervisor's office (5). How to take an appeal. How to take a dispute. Definition of terms used in connection with appeals and disputes. 3. The board of review. a. Location. b. Duties. 4. Secretary of Agriculture. a. Relation to the inspection and grading work. 5. Relations between inspectors, supervisors, board of review, and Secretary of Agriculture. (Ex- plained by use of diagram on blackboard.) III. Bases upon which the grain grades are established. In this part of the course the students are assigned references to report on and discuss before the class. The references used are listed below in about the order in which they are taken up by the class. 1. U. S. Dept. Agr. Bui. 328. Milling and baking tests of wheat containing admixtures of rye, corn cockle, kinghead, and vetch. 2. U. S. Dept. Agr., Bur. Plant Indus. Bui. 100. Garlicky wheat. 3. Ind. Agr. Expt. Sta. Bui. 176. Wild garlic and its eradication. 4. U. S. Dept. Agr. Farmers' Bui. 610. Wild onion : Methods of eradication. 5. U. S. Dept. Agr. Bui. 455. The drying for milling purposes of damp and garlicky wheat. 6. U. S. Dept. Agr. Farmers' Bui. 919. Methods of handling dockage. 7. U. S. Dept. Agr. Bui. 557. A comparison of several classes of American wheats and a consideration of some factors influencing quality. 8. Bureau of Markets Service and Regulatory Announcements 54. How hard red winter wheat is grading under Federal standards. 9. U. S. Dept. Agr. Bui. 788. Moisture in wheat and mill products. 10. U. S. Dept. Agr. Bui. 751. Experiments in the digestibility of wheat bran in a diet without wheat flour. 11. Utah Agr. Expt. Sta. Bui. 103. Milling qualities of wheat. 12. Canadian Dept. Agr. Bui. 57. Quality in wheat. 13. U. S. Dept. Agr. Bui. 48. The shrinkage of corn while in cars in transit. Journal of the American Society of Agronomy. Plate 8. Fig. 2. Closer view of some of the grain-grading apparatus. WENTZ : COURSE IN GRAIN GRADING. 20I 14. U. S. Dept. Agr. Bui. 725. A preliminary study of the bleaching of oats with sulphur dioxid. 15. U. S. Dept. Agr., Bur. Plant. Indus. Cir. 74. The sulphur bleaching of commercial oats and barley. IV. Laboratory practice. A large number of samples of the different grains are obtained from various sources, such as the local experiment station, feed stores, and farmers of the State, and assigned to the students. The first laboratory period is used in demonstrating the use of the grain probe in obtaining samples and the apparatus used in grading, and in dis- cussing various phases of the different operations. In the following laboratory periods the students are furnished with direction sheets to be followed in grading the assigned samples, and with the use of the small handbook, " Official Grain Standards" (8), and type samples, the students are able to go ahead with a reasonable amount of help from the instructor. The following are copies of the direction sheets used. Directions for Grading Wheat. 1. Give your sample a laboratory number. 2. Make moisture test, using sample contained in air-tight can (see Manual, P- 33-36). For all other determinations that portion of the sample con- tained in the cloth bag is used. 3. Determine odor, onions, and garlic, and live weevils or other insects injurious to stored grain. REPORT ON WHEAT GRADING TESTS. Sample number Percent of moisture . . . Odor Percent of dockage Weight per bushel Damaged kernels: Total Heat damaged . . . . Foreign material: Total Other than cereals . Class or subclass Wheats of other classes Grade Remarks 4. Divide sample down to about 1,000 grams (see Manual, p. 31, 32). 5. Determine dockage, using the 1,000 gram sample (see Manual, p. 14 and 37-42). 6. Determine test weight per bushel, using the dockage-free wheat obtained in 5 (see Manual, p. 13 and 43). 202 .JOURNAL OF THE AMERICAN SOCIETY OF AGROT OMY. 7. Divide the sample down to three portions, A, B, and ^containing 25 to 65 grams each. 8. Using portion A, determine total damaged kernels, heat-aamaged kernels, total foreign material, and foreign material other tha 1 cereal grains. 9. Using portion B, determine the class and subclass into which the sample should be placed by analyzing for color and texture. 10. Using portion C determine wheats of other classes. DIRECTIONS FOR GRADING CORN. 1. Give your sample a laboratory number. 2. Make moisture test, using sample contained in air-tight can (see Manual, P- 33-3^)- For all other determinations use that portion of the sample contained in the cloth bag. 3. Determine odor, live weevil or other insects injurious to stored grain, and quality. 4. Divide sample down to about 1,000 grams. 5. Determine weight per bushel, using the 1,000-gram portion. 6. Divide down to 240 to 260 grams. 7. Determine foreign material and cracked corn (see Manual, p. 21). 8. Using the cleaned sample, determine total damaged and heat-damaged kernels. 9. Divide down to about 100 grams. 10. Determine color, using the 100-gram portion. REPORT ON CORN GRADING TESTS. Sample number Percent of moisture Odor . . . • Weight per bushel Foreign material and cracked corn Damaged kernels: Total Heat damaged Color or class Grade Remarks DIRECTIONS FOR GRADING OATS. 1. Give your sample a laboratory number. 2. Make moisture test, using sample contained in air-tight can. For all other determinations use that portion of the sample contained in the cloth bag. 3. Determine condition and general appearance. 4. Divide sample down to a portion of 500 to 600 grams. 5. Determine weight per bushel, using the 500-gram portion. 6. Divide down to three portions, A, B, and C, each containing 25 to 65 grams. 7. Using portion A, determine sound cultivated oats, heat-damaged (oats or other grains), and foreign materials. When a sample contains an unusual amount of foreign material, make determinations for this factor on not less than 250 grams of the original sample, using the buckwheat sieve for removing the seeds and dirt, WENTZ I COURSE IN GRAIN GRADING. 203 recovering aViy small pin oats that may pass thru with the dirt by re- screening, then hand-pick the sample for any remaining foreign material. 8. Using portion B, determine percentage of wild oats. 9. Using porriori 1 C, determine percentage of other colors of cultivated and wild oats REPORT ON OAT GRADING TESTS. Sample number Percent of moisture Condition and general appearance . . . Weight per bushel Sound cultivated oats Heat damaged (oats or other grains) . Foreign material Wild oats Other colors cultivated and wild oats Color or class Grade Remarks After completing the course as here outlined, the student has the fundamental principles of grain judging and will have a tendency in his later experiences in judging grains to lay stress upon facts brought out in such a course. The instructor has a chance to connect the vari- ous factors taken into consideration in the grade requirements with the practical value of the grains, and formulate in the students' minds the most practical score card to be used in judging grains at a grain show, on the farm, or in commerce. In the course as it is now being conducted much interest and value has been added by a lecture by Mr. Harold Anderson, Grain Super- visor of this district, on " The Duties of the Grain Supervisor and his Relations to the Farmer," and a trip to Baltimore to inspect the supervisor's laboratory, the Chamber of Commerce trading floor and inspection laboratory, and a large grain elevator. Another factor which gives variety and interest to the course and seemed to be well worth while was the showing of the following motion picture films loaned by the United States Department of Agriculture : 1. The Story of Wheat in the Pacific Northwest, 2 reels. 2. Wheat : Transportation and Storage, 3 reels. 3. Wheat Grading Under Federal Supervision, 1 reel. Literature Cited. 1. Brown, Ralph H. The farmer and federal grain supervision. In Year- book of U. S. Dept. Agr. for 1918, p. 335-346. 1919. 2. Bureau of Markets. Rules and regulations of the Secretary of Agricul- ture under the United State Grain Standards Act of August 11, 1916. U. S. Dept. Agr., Off. Sec. Circ. 70. 1916. 204 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. 3. . Official grain standards of United States for wheat and shelled corn. U. S. Dept. Agr., Service and Regulatory Announcements No. 33. 1918. 4. . Official grain standards of United States for oats. U. S. Dept. Agr,, Service and Regulatory Announcements No. 46. 1919. 5. . The country grain dealer and Federal grain supervision. U. S. Dept. Agr., Service and Regulatory Announcements No. 47. 1919. 6. . How to take an appeal. U. S. Dept. Agr., Service and Regulatory Announcements No. 52. 1919. 7. Miles, R. T. History and purpose of grain grading. In Report of Con- ference of Teachers of Farm Crops in a Number of Agricultural Colleges, with Federal Grain Supervision Officials of the Bureau of Markets, U. S. Department of Agriculture. 8. U. S. Grain Standards, Form No. 90. Handbook Official Grain Standards for Wheat, Shelled Corn, and Oats. HARLAN : SMOOTH-AWNED BARLEYS. 205 SMOOTH-AWNED BARLEYS. 1 Harry V. Harlan. Barleys with smooth awns or partially smooth awns are so little known that, altho the term " smooth-awned " is descriptive, it rarely conveys the true meaning. This is because hooded and awnless varie- ties are frequently knows by such names as "smooth" and "bald." The hooded varieties are widely distributed, the Nepal being found in field culture in every barley-growing State, but awnless varieties are little known. In the hooded varieties, the awn is replaced by a three- pronged hood that extends slightly beyond the kernel, the prongs being turned backward. This hood is not harsh in structure as is the awn. The so-called awnless varieties have no awns on the lateral florets, but usually have short awns on the central florets. The edges of the awns of the com- mon rough-awned barleys are armed with closely-set projecting teeth which point toward the tip (fig. 15, b) . The largest teeth are at the base of the awn, the teeth gradually decreasing in size until near the tip they are so small that the awn feels only slightly rough. In the smooth-awned sorts the teeth are almost entirely wanting (fig. 15, a). For all practical pur- poses, they are eliminated, altho all smooth-awned barleys are not equally smooth. The large teeth near the base are the first to disappear. Some varieties are smooth for two-thirds of the distance from the base to the tip, Fig 15. Section of (a) smooth . . . tee an( l (&) rough awns of barley while others are smooth for four- (after Regel). fifths or more of the length. As the teeth on the upper third of the awn are not objectionable, all barleys of this class are essentially smooth-awned. These awns are so smooth that they may be pulled across the face in either direction without roughness being apparent except at the tip. 1 Contribution from the Office of Cereal Investigations, Bureau of Plant Industry, United States Department of Agriculture, Washington, D. C. Re- ceived for publication April 5, 1920. 206 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. The economic advantage of smooth-awned barleys is considerable. The rough awns have limited the production of this grain in many ways. It is a disagreeable crop to handle during harvesting and thrashing and for this reason farmers often object to growing barley. Growers frequently have difficulty in exchanging labor at thrashing time, if their proportion of barley is large. The straw frequently causes sore mouths in animals, the pieces of awns piercing the mem- branes and being held there by the numerous curved teeth. When sheep are fed on barley, the awns work into the wool. In the West, barley is used much for hay and would be more widely grown for this purpose if it were not for the rough awns. Yield being equal, every farmer would prefer a smooth-awned variety, and if a high- yielding one can be found, it will doubtless result in an increased acreage of the crop. Smooth-awned barleys are not new. They have been known in European botanical collections for many years. Koernicke described the leiorrhynchum form in 1882. In the same paper he described medic urn and persicum, smooth-awned 2-rowed sorts that are found in mixtures of barley from Asia Minor. Robert Regel in Russia has done very extensive work with smooth- awned varieties and published a monograph on them in 1909. He reported a considerable number of such barleys from Russia. Various Russian experiment stations have tested these varieties. So far as the writer knows, no pure culture of a smooth-awned barley has been grown commercially. Little interest has been shown in smooth-awned barleys in North America. Smooth-awned stocks have been limited and the work done in Russia has not attracted the attention it deserved. The Office of Cereal Investigation of the United States Department of Agriculture has studied these forms for several years. A large part of the work has been in cooperation with the Minnesota Agricultural Experiment Station, where the writer became the representative of the United States Department of Agriculture in some cooperative barley investi- gations in 1909. At harvest time that year, a plant with partially smooth awns was found in a plat of Hanna barley, many strains of which show a tendency toward smoothness. The plant found had the teeth missing from the lower one-third of the awns. It was the inten- tion to use the progeny of this plant in breeding operations to obtain a smooth-awned form, but other forms with fewer teeth were found in 1910. One plant was found in the Prinsess variety and one in a barley from southern Russia. In 191 1 and 1912, crosses were made which included very smooth H. leiorrhynchum parents. One strain of HARLAN : SMOOTH-AWNED BARLEYS. 20/ leiorrhynchum, while very smooth, had a weak straw and was replaced in most of the later crosses by a leiorrhynchum form received in 191 2 from Russia. This importation was sent to the Office of Foreign Seed and Plant Introduction by F. H. Meyer from Taganrog in south- ern Russia. The Taganrog barley was not as smooth-awned as that used in 191 1 and 1912, but it was much superior in every other way. The F l generation of several crosses w r as grown in the greenhouse at Washington during the winter of 1912-1913. The seed from this was sown in May, 1913, at the Minnesota station. Counts of the progeny of the F 2 generation showed smoothness to be recessive. The breeding of smooth-awned sorts w r as thus found to be a very simple matter and several hundred plants were selected from crosses already made. The rough-awned parents included Manchuria, Mariout, and Coast. Many other crosses were made and a considerable number were in the F 1 generation at this time. In 191 3, seed of the Russian smooth-awned parent was sent to California, Idaho, and Michigan. Small lots of the Taganrog barley and lines obtained from various crosses were sent out in succeeding years until by 191 7 smooth-awned strains were in the hands of ex- perimenters from New York to California and some seed had been sent to Canada. In the meantime, a smooth-awned barley had been received thru the Office of Foreign Seed and Plant Introduction, from Dr. Trabut of Algeria, in July, 1910. This, like the Taganrog barley, was a black 6-rowed form. Only a small quantity of seed was received and this was sent to California and Oregon under the name of Black Algerian (C. I. No. 708). The writer received none of the seed and did not learn of the existence of this variety until it had been grown for several years. It does not appear to be as smooth as the strains grown at St. Paul. Other smooth-awned forms have been introduced thru the Office of Foreign Seed and Plant Introduction. These include two 2-rowed forms from Asia Minor, one white and the other black. Neither of these has been used in crosses. More recently a smooth-awned white 6-rowed form was received from Robert Regel. It is not known whether or not independent importations have been made. Partially smooth sorts have been isolated from our common barleys by the writer and doubtless by others. None of these that has been included in the experiments reported has been as smooth as the better strains of H. leiorrhynchum. The breeding operations have been conducted at St. Paul, Minn. ; Arlington Farm, Va. ; Chico, Cal. ; Aberdeen, Idaho ; and Moro. 208 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Oregon. A large number of crosses have been made, embracing a wide range of rough-awned parents. Several hundred strains have been isolated and studied. It has been found entirely practicable to secure segregates from all matings with the basal two-thirds of the awn smooth, no matter how rough the awn of the rough-awned parent may be. As we have no conception of the number of factors in a strain, it is hardly possible to say that the teeth of the awns of a variety can be removed by crossing. Types similar to the rough- awned parent except for the absence of teeth on the awns may be isolated readily, but it is extremely unlikely that the variety is not otherwise changed as well. As barleys of this type have not yet been tested under farm condi- tions, and as most of the yields obtained are nursery yields, it is not thought worth while to present records of performance at this time. From the results to date, however, there seems to be no reason why high-yielding smooth-awned barleys can not be produced. The origi- nal importations of H. leiorrhynchum have yielded as well as the average importations of rough-awned barleys. Smooth-awned crosses have yielded well in the nurseries at various places, and these yields have been maintained over a series of years. On the other hand, the absence of such varieties in commercial cultures in Europe and Asia indicates a weakness of some sort. Such forms have been known for many years. They must have been observed long ago by peasants in Asia Minor and southern Russia. Why has not such an attractive feature been utilized ? At present we can not even guess, for so far the yields indicate that smooth-awned barleys may be made to give satisfactory harvests. * Smooth-awned barleys are still in the experimental stage. Several high-yielding strains adapted to different climatic conditions are ready for increase to larger plats and to field culture. Whether they can compete with rough-awned varieties remains to be determined. AGRONOMIC AFFAIRS. 209 AGRONOMIC AFFAIRS. ANNUAL MEETING OF THE SOCIETY. The thirteenth annual meeting of the American Society of Agron- omy will be held at Springfield, Mass., October 18 and 19, 1920, in connection with the annual meeting of the Association of Land Grant Colleges. NOTES AND NEWS. W. E. Ayres, formerly assistant agronomist at the Arkansas sta- tion, is now agronomist at the Delta Substation, Stoneville, Miss. M. A. Beeson, head of the department of agronomy at the Okla- homa college and station, resigned July 1 to engage in commercial work. Manley Champlin, for several years associate agronomist of the South Dakota college and station and more recently extension agron- omist, has resigned to become senior field husbandman in the Univer- sity of Saskatchewan. R. W. Clothier, for the past several years with the Federal Office of Farm Management, is now president of the New Mexico Agri- cultural College. H. R. Cox is now extension specialist in crops and soils in New Jersey. W. H. Darst, formerly associate professor of farm crops in the Pennsylvania college, has resigned to accept a position with the North Carolina A. & M. College. Dr. Spright Dowell, state superintendent of education, has been elected president of the Alabama A. & M. College, effective July I. The former president, Dr. C. C. Thach, has been made president emeritus. Arthur T. Evans, formerly professor of botany and dean of Huron College, is now associate agronomist of the South Dakota college and station. Ernest M. Fergus, formerly instructor in farm crops at Purdue University, is now connected with the department of agronomy in the University of Kentucky. • 4 2 10 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. D. S. Fox, assistant professor of agronomy at the Pennsylvania college, has resigned, effective July i. E. F. Gaines, plant breeder at the Washington station, is on leave of absence beginning September I, for postgraduate study at Bussey Institution. H. H. Laude, formerly agronomist at the Texas station, is now in charge of cooperative experiments at the Kansas station. He is as- sisted by N. E. Dale, who has succeeded Bruce S. Wilson. Guy R. McDole, assistant soil chemist at the Minnesota agricul- tural college and experiment station, has accepted the position of associate professor of agronomy and soil technologist at the Univer- sity of Idaho. Paul V. Maris, for four years state leader of county agents in Ore- gon, has been made director of extension in that state. G. K. Middleton, associate professor of agronomy at the North Carolina college, has resigned to take up agricultural work at Kai- pang, Hainan Province, China. J. R. Nevius is now instructor in farm crops in the Oregon college. J. S. Owens, formerly extension agronomist at the Pennsylvania college, is now on the field staff of the eastern bureau of the National Lime Association. R. C. Parker, formerly county agent in Suffolk County, L. I., now represents the National Lime Association in New England tmd east- ern New York, with headquarters at Springfield, Mass. J. L. Robinson, formerly assistant agronomist at the Wyoming col- lege and station, is now director of cooperative experiments in the farm crops department of Iowa State College. Karl Sax, formerly engaged in plant breeding at Riverbank Lab- oratories, Geneva, 111., has been appointed biologist at the Maine sta- tion in charge of plant breeding and has entered on his new duties. Robert Stewart, for the past several years professor of soil fer- tility at the University of Illinois, is now dean of the college of agri- culture of the University of Nevada. Prof. Samuel Mills Tracy, formerly director of the Mississippi Agricultural Experiment Station, and for the past 23 years connected with forage-crop investigations of the Federal Bureau of Plant In- dustry, died at Laurel, Miss., September 4, 1920, aged 73 years. Pro- fessor Tracy was born in Hartford, Vt., April 30, 1847. He grad- uated from the Michigan Agricultural College in 1868 and received the degree of M.S. from the same institution- in 1871. During the AGRONOMIC AFFAIRS. 21 I civil war he was a private in Company A, 41st Regiment, Wisconsin Volunteers. He was professor of botany and horticulture at the State University of Missouri from 1877 to 1887, and director of the Mississippi Agricultural Experiment Station from 1887 to 1897. Since 1897 he has been engaged in a study of southern forage crops for the Bureau of Plant Industry, and during much of this time he has conducted an experiment station on his farm at Biloxi, Miss. During the past two years he has had charge of forage crop work at the McNeill, Miss., Substation. In addition to his forage crop inves- tigations, Professor Tracy did much botanical work in connection with the flowering plants and fungi of the Southern States. Collec- tions of his material are in many of the largest herbaria, and his own private collection is now at the Texas Agricultural College. Pro- fessor Tracy is the author of many bulletins of the U. S. Department of Agriculture and of the Mississippi Agricultural Experiment Sta- tion. He was a man of charming personality and his death is a dis- tinct loss to his many friends and coworkers. H. J. Webber has resigned as director of the California Agricul- tural Experiment Station to engage in commercial work, he now being connected with the Coker Seed Farms, Hartsville, S. C. He has been succeeded as director by Dr. C. M. Haring, formerly professor of veterinary science. James Wilson, formerly professor of agriculture and director of the Iowa station, and Secretary of Agriculture in the cabinets of Presidents McKinley, Roosevelt, and Taft for the 16 years from 1897 to 1913, died at Traer, Iowa, August 26, at the age of 85 years. The organization of the National Research Counsel for the year beginning July 1, 1920, is as follows: Chairman, H. A. Bumstead, professor of physics at Yale University ; 1st vice-chairman, C. D. W T alcott, secretary of the Smithsonian Institution ; 2nd vice-chair- man, Gano Dunn, president of the J. G. White Engineering Corpora- tion ; permanent secretary, Vernon Kellogg, formerly professor of zoology at Stanford University. CONFERENCE ON ELEMENTARY SOIL TEACHING. The following report of the Conference on Elementary Soil Teach- ing has been received from the Secretary, Prof. P. E. Karraker, of the University of Kentucky. A very enjoyable and, it is believed, profitable meeting of soils in- structors in the State agricultural colleges was held at the College of 212 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Agriculture, University of Kentucky, June 23, 24, and 25, 1920. The purpose of the conference was a general discussion of the teaching of elementary soils in the State agricultural colleges. Attention was specially directed to the securing of greater uniformity in presenting this subject in the various colleges and to the determination of what should properly constitute the laboratory part of the work. The con- ference was called in response to a widespread feeling of need for such a meeting. Mention, however, should be made in particular of the very large part Dr. H. O. Buckman, of Cornell University, had in bringing it about. A semi-permanent organization was effected by retaining the chair- man, Prof. M. F. Miller, of the University of Missouri, and the sec- retary in their positions to facilitate such future action as may seem desirable. It was the feeling that meetings, such as this, of in- structors in the various lines of work in the State agricultural col- leges are well worth while and should be encouraged. There was no feeling of finality in the results achieved by the con- ference and consideration and criticism by teachers not present is earnestly invited. It is believed, however, that the getting together of this group of instructors in discussion of their work and the spe- cific recommendations resulting therefrom are a considerable advance toward the better teaching of soils. The authorized report of the conference is as follows : Representatives of the soil or agronomy departments of sixteen State agricultural colleges, assembled in conference at the University of Kentucky, June 23-25, 1920, agree upon the following recom- mendations : " The first college course in soils should be a uniform general course known as the ' Principles of Soil Management.' This course should be required of all agricultural students and should carry ap- proximately five semester hours credit. "This course should deal largely with the scientific principles un- derlying the successful management of soils in general, with such practical applications as conditions demand. " The minimum prerequisites of this uniform course should be one year of general inorganic chemistry, one term of general geology, and either high school or college physics. "The subject matter of the course should be presented by well correlated lecture, recitation and -laboratory work. At least three- fifths of the time should be utilized in lecture and recitation. Wher- ever practicable the work should be given in the Sophomore year. It is desirable that a standard text book be used. AGRONOMIC AFFAIRS. 213 " The laboratory exercises covering one semester hour's time should conform as nearly as possible to the following general outline ; where two laboratory periods are used the quantitative method should be applied to this outline : Rocks, minerals and weathering (one or more periods). The soil particles and soil class (one period). Volume weight determinations, either in the field or laboratory, with calcula- tions (an optional one-period exercise). f Peat, muck or a highly organic soil. Study of organic matter -I Maximum water as affected by organic matter. [ Estimation of organic matter, etc. Study of soil structure either in field or laboratory, including a specific study of colloids. (The study of organic matter and structure shall cover one or more periods.) Soil moisture. (Two or more exercises as conditions make necessary.) Optimum moisture and moisture determinations. Moisture determinations on field soil under different treatments. Maximum water capacity. Estimation of soil moisture. Calculations, etc. Soil heat (an optional one-period exercise.) Temperature of field soil. Heat conductivity, etc. Absorption of nutrients by the soil. (An optional one-period exercise.) Soil reaction studies — acidity and alkalinity (one or more exercises). Fertilizer materials and lime (two or more exercises). Interpretation of soil survey reports and maps (one period). At least one assigned problem in soil management to be presented and discussed at any time that may be convenient. Note: Two or more field trips to be given after any of the following exer- cises: (1) soil class; (2) organic matter and structure; (3) soil moisture; (4) fertilizer materials and lime. " The advantages of such a course are as follows : " 1. The student is able to obtain in one course a survey of the entire subject. " 2. The course will make possible the preparation of standard texts and illustrative materials and standard laboratory equipment. " 3. Such a uniform course will facilitate the transfer of credits from one institution to another." The representation at the conference was as follows : Firman E. Bear, Ohio State University; H. O. Buckman, Cornell University; Geo. A. Crabb, Georgia State College of Agriculture ; L. F. Gieseker, Montana State College; T. B. Hutcheson, Virginia Polytechnic Institute; P. E. Karraker, University of Kentucky; 214 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. A. F. Kidder, Louisiana State University; H. A. D. Leggett, University of Vermont ; R. B. Lowry, University of Tennessee; A. G. McCall, University of Maryland ; C. Ernest Millar, Michigan Agricultural College; M. F. Miller, University of Missouri; Geo. Roberts, University of Kentucky ; R. S. Smith, University of Illinois; Robert Stewart, University of Nevada ; D. C. Wimer, Pennsylvania State College; Casper A. Wood, Agricultural and Mechanical College of Texas. CONFERENCE OF WESTERN AGRONOMISTS. The following report of the Conference of Western Agronomists has been received from Roland McKee of the United States Depart- ment of Agriculture. The fourth annual conference of agronomists of the Rocky Moun- tain and Pacific Coast States was held at the University of California, Berkeley, June 8 and 9, and at the California Agricultural Experi- ment Station, Davis, Cal., June 10. The meeting was well attended, six States and the U. S. Department of Agriculture being repre- sented. Prof. J. W. Gilmore of the University of California was chairman. The program consisted of discussions of the following subjects: Problems of Power in Tillage and Harvest; Soil Problems Relating to Crop Production ; Farm Crop Diseases and Treatments ; Farm Crops and Seed Production and Utilization; Problems Relating to Teaching and Leadership. On the evening of June 8 three reels of motion pictures showing modern operations in wheat production were presented by J. A. Clark of the U. S. Department of Agriculture. The problems of power in tillage and harvest were presented by Prof. W. J. Gilmore of the Oregon Agricultural College. The dis- cussion indicated that much progress has been made in determining the place of tractor power on the farm, but it also seemed evident that varying local conditions make the application and use of tractors a problem for local consideration. Dr. F. S. Harris, director of the Utah Agricultural Experiment Station, presented in outline form the alkali soil problems as related to crop production. The great importance of this subject to the agri- culture of the West was emphasized and the complexity of the prob- AGRONOMIC AFFAIRS. 215 lems indicated. In spite of the handicaps, Doctor Harris is very opti- mistic of the possibilities of much good work being done. Emphasis was laid on the necessity of more knowledge on the subject so that the future possibilities and permanency of crop production on lands that may be developed under irrigation may be more accurately foretold. The subject of farm crop diseases and treatments was presented by Prof. W. W. Mackie of the University of California. In connec- tion with the treatment of seed grain for loose smut by the use of formaldehyde, recent work done at the University of California in cooperation with the U. S. Department of Agriculture has indicated that much damage may be done by allowing the seed to dry after treating or by sowing in dry soil. This is due to the formation of paraformaldehyde, which kills the germinating seed. D. E. Stephens, superintendent of the Sherman County Branch Station, Moro, Oreg., and Prof. F. J. Sievers of Washington State College led in a discussion of some problems in soil tillage and rota- tion in cereals. Mr. Stephens had found in his work that anything that lengthens the season of maturing of wheat increases " yellow berry." The most common factors in this connection are nitrates, organic matter, and moisture. He has secured no difference in yields of grain following 5 and 8 inch depths of plowing. Similar results have been obtained at Pullman, Wash. Fall disking has been of no advantage. Early spring plowing has given better results than later spring plowing. The amount of moisture in the fall has been prac- tically the same on fallow plats receiving good cultivation, poor cul- tivation, and no cultivation. Many good points were brought out in these and other discussions in which the work being done at various institutions was presented. The last day of the conference was spent at the University Farm, Davis, Cal., where opportunity was afforded for seeing the field ex- periments conducted at that point. The conference in 1921 will be held in Arizona, the exact date to be determined by the directing committee, consisting of Prof. G. E. Thompson of the University of Arizona, Prof. J. W. Gilmore of the University of California, and Roland McKee of the U. S. Depart- ment of Agriculture. 2l6 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. MEMBERSHIP CHANGES. The membership reported in the May Journal was 538. Since then, 13 new members have been added, 3 have been reinstated, 2 have resigned, and 2 have died, making a net gain of 12 and a total present membership of 550. The names and addresses of the new and reinstated members, the names of those resigned and deceased, and such changes of address as have been noted, are as follows : New Members. Branstetter, B. B., University of Missouri, Columbia, Mo. Friant, R. J., Delaware College, Newark, Del. Green, W. J., Experiment Station, Agana, Guam. Hadley, F. E., ioi California St., San Francisco, Cal. Jensen, Irving, Agr. Expt. Sta., Logan, Utah. Kirkpatrick, Roy T., University of Missouri, Columbia, Mo. Meyers, M. T., 58 W. Frambes Ave., Columbus, Ohio. Parker, R. C, Box 505, Riverhead, N. Y. Powers, Wm. H., Agricultural College, Brookings, S. Dak. Seaton, Jerome P., Glencarlyn, Va. Rohde, W. C, Agriculturist, The Barrett Co., Baltimore, Md. Tompkins, J. F., Burdette, Ark. Zahnley, J. W., Kansas State Agr. College, Manhattan, Kans. Members Reinstated. Freeman, George F., Societe Sultanienne d'Agriculture, Cairo, Egypt. Scott, Herschel, Betteravia, Cal. Snyder, R. M., East Lansing, Mich. Members Deceased. S. M. Tracy, H. Foley Tuttle. Members Resigned. F. S. Hagy, C. K. McClelland. Changes of Address. Biggar, H. H., The Dakota Farmer, Aberdeen, S. Dak. Blair, R. E., Box 641, Porterville, Cal. Bracken, John, Agricultural College, Winnipeg, Manitoba. Breithaupt, L. R., Ontario, Oregon. Buie, T. S., Agr. Expt. Sta., Clemson College, S. C. Champlin, Manley, University of Saskatchewan, Saskatoon, Sask. Cromer, C. O., Daleville, Ind. Darst, W. H., A. & M. College, West Raleigh, N. C. Duley, F. L., 27 Allen Place, Columbia, Mo. Dunlavy, Henry E., Italy, Texas. JOURNAL OF THE American Society of Agronomy Vol. 12. November-December, 1920. Nos. 8-9 THE AGRONOMIST'S PART IN THE WORLD'S FOOD SUPPLY. 1 F. S. Harris. The welfare of mankind is intimately bound up with the world's food supply. Not that man can "live by bread alone," but he is unable to devote himself to the higher phases of an advancing civili- zation if he is conscious of the gnawings of hunger. Since the short- age in various food products during the war, people generally have taken a much keener interest in the whole question of food supply. The old statement that " we never miss the water till the well runs dry " is here exemplified. So long as the grocer had plenty of flour and sugar, most people considered the supply in much the same way as they considered the supply of air. The only worry was to find money with which to purchase needed articles. When it became necessary to go to a dozen stores before being able to buy any sugar, and then only a pound or two ; when the meat allow- ance was restricted ; and when white flour had to be supplemented by all kinds of substitutes, then people began to realize that the supply of food might not be inexhaustible. The shortage of food during the war has been a good lesson for the people of the United States. It has taught them what some of the people of Asia have been so often forced by famine to realize, namely, that food can be had only when a supply is available and that this supply may at times be far short of actual needs. Conditions during the war were of course unusual ; we hope they will never recur. I do not at this time desire to consider the food shortage due to the war but rather the whole food situation as it is likely to affect man- 1 Presidential address presented at the thirteenth annual meeting of the American Society of Agronomy, Springfield, Mass., October 18, 1920. 217 2l8 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. kind in the future as the population of the earth increases. There will be of course temporary local short-time food shortages due to unfavorable seasons, wars, or other unusual conditions. These situa- tions will have to be met as best they can at the time. The thing to which I should like to direct attention at present is not this temporary or local condition of famine but rather the means by which people may be fed when the world becomes much more populous than it now is. Having an earth, the best land of which is already producing crops without any great surplus, how is it going to be possible for nations to grow, cities to be built, and civilization to advance? Is there a limit to the number of people for whom the earth can supply food, or can the increase go on indefinitely? As a small boy I remember going through what seemed to me to be an immense forest with a man who said it contained enough timber to last the whole United States for a thousand years. Later, when I became old enough to make the calculation, I found that this par- ticular body of timber would not furnish America's needs for a single year. In the early days of the settlement of the West, many who saw the large rivers made the statement that the water of these streams could never be exhausted by irrigation. The supply was said to be limitless. Experience has shown that the water of many of these streams was exhausted before more than a fraction of the land adja- cent to them could be served. Thus, all things have their limits. There is a limit to the number of people a given area of land can sustain and, as the area of land is practically constant, there must be a limit to the number of people that can be fed. The number of course depends entirely on how fully the resources of the earth are utilized. It is possible greatly to increase production. I wish later to call attention to the methods by which the agronomist may assist in accomplishing this end. I am not an alarmist. I do not wish to appear as one who is trying to stir people up unnecessarily. I should not like even to take the responsibility assumed by Sir William Crookes, who, in his presiden- tial address before the British Association for the Advancement of Science in 1898, set a date when the shortage of food would begin to be felt. I do not believe that sufficient data are available for any- one to be so definite. A few facts, however, may be used to help in clarifying our minds on the subject. It is well known that the population of all important countries of the world is gradually increasing. During the no years from 1800 HARRIS : THE WORLD'S FOOD SUPPLY. to 1910 the population of the world increased from 640,000,000 to 1,600,000,000, or an increase of 152 percent. Only a few generations ago there were vast continents of unsettled fertile land waiting to absorb the overflew from the populous parts of the world. There are still many large tracts of land that are not settled, but it is obvious to all who have made a study of the subject that the better lands are rapidly being put under cultivation, and only the more remote and more unfavorable areas remain. This does not say that there is not still available much excellent land ; but let us consider the United States as an example. In 1790 the population of the entire country was 3,929,214; by 1840 it had reached 17,059,453; while the 1920 census shows it to be more than 105,000,000. A century ago only the east coast was settled; the great heart of the agricultural land had not been touched. The ra- pidity of settlement of the Central States is indicated by the fact that between 1800 and 1820 the population of Ohio, Indiana, Illinois, Michigan, Wisconsin, and Iowa increased from 50,240 to 792,719, and by 1840 it had reached 2,967,840. Today the entire country has been thoroly explored and the better land has been producing for nearly a generation. In order to see just how our production and consumption have bal- anced during the last three score and ten years — the allotted time of man — let us examine the figures for wheat, probably our best index crop. Table i. — Average wheat production in and export from the United States by 10-year periods from 1849 to 1919. Year or decades. Average annual pro- duction. Average annual exports. Percent of total crop exported. Bushels. Bushels. 1849 100,486,000 7,535.901 7-5 1859-1868 190,395.750 21,475,072 11.3 1869-1878 285,951,600 73.634,732 25-7 1879-1888 446,587,600 133.703,079 29.9 1889-1898 495,184,800 165,377.944 33-3 1899-1908 651,643,800 152,533,604 23-4 1909-1918 754,471,400 172,400,807 22.8 Average of only four years, 1859, 1866, 1867, 1868. In 1849 we produced approximately 100,000,000 bushels of wheat in the country, only 7.5 percent of which was exported. With the rapid settlement of the West, production rose till during the decade from 1879 to X S88 it reached 446,587,600 bushels, 33.3 percent of which was exported. Thus new productive land was brought under 2 20 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. cultivation much faster proportionately than population increased. After this time, however, the population so gained on production that during the next decade only 23.4 percent of the wheat produced was exported, and during the ten years from 1909 to 191 8 the exports averaged 22.8 percent of the production. This figure was much in- ci eased by enlarged exports due to the war. During the years imme- diately preceding the war the exportation of wheat had almost ceased. In 1800, 80.4 percent of our exports consisted of agricultural prod- ucts, whereas in 1910 the percentage had dropped to 50.9 percent. These figures are significant because they show that, even in a country like the United States where the area and resources seem to be almost limitless, it will not long be possible to continue to feed other than our own increasing population. A condition that helps to bring this about is the rapidly increasing proportion of our city-dwelling population. In 1820 only 4.93 per- cent were urban. In 1880 this had reached 29.5 percent, leaving still 70.5 percent rural; ten years later 36.1 percent were urban and 63.9 percent rural; in 1900 40.5 percent were urban; and by 1910, 46.3 percent were urban and only 53.7 percent rural. The 1920 census shows that more than half of the people of the United States are living in cities. With a condition of this kind the food situation is likely to become more acute than where most of the population live on the farm where they can more quickly influence the rate of food production. With the growth of many large cities and with the complex systems of modern transportation and exchange, the food question tends more and more to become a single whole-world problem rather than nu- merous small local problems affecting the smaller communities. With our modern systems unobstructed by war, we shall probably never again have such devastating local famines as were so common in past generations in India, China, and Russia during years when there was an abundance in other parts of the world. The situation as it appears to me is this : We live in a world with an increasing population. This increase can not expand indefinitely to fertile unoccupied lands, for these lands are becoming scarce. The food supply must be increased as fast as the population increases, for food supply is the chief limiting factor in population growth. There is no immediate cause for alarm, but it is the duty of scien- tists and statesmen to look to the future. We must not be content to be like Sam the negro, who took his stove to his boss and offered it for sale for a fraction of its value. On being asked if he would not HARRIS : THE WORLD'S FOOD SUPPLY. 221 need it next winter he said he would but that winter was three months away while the circus was tomorrow ! Satisfying the needs of today is not sufficient. We must maintain a forward-looking attitude. It is impossible to make large increases in production quickly ; years of preparation and work will be required to do anything of permanent value. An adequate solution of the world's food problem can be made only by deliberate planning. All factors involved must be considered and a world-wide program of work initiated, for the world is now a unit in production and in con- sumption. The problem will involve a great variety of business and scientific interests. Credit, transportation, manufacturing, and mechanics must all be called on to do their part. What we are now most inter- ested in, however, is the contribution of the agronomist. What is his part in the world's food problem? An examination of the question indicates that his part is a large one. While it is not entirely clear just what is included under the word " agronomy," the general understanding is that it has to do with anything affecting crop production and, as the food supply is in the last analysis a question of crop production, it would appear that the agronomist has a great responsibility in seeing that the people of the world do not want for something to eat. Let us see what means he has available to meet this responsibility. We have already shown that the increasing population will call for increased production. This increase can be met in just two ways : First, by extending the producing area, and second, by increasing the acre yield of the present cultivated area. The method of enlarging the agricultural area will be discussed under the following four headings : ( i ) Increasing the irrigated area ; (2) extending dry farming; (3) drainage of wet lands; and (4) reclamation of alkali lands. Of course there are uncultivated lands in the world that will not require any of the methods of reclamation mentioned above to make them productive. They may be inaccessible, or for some economic reason it may not pay to cultivate them even tho they are fertile. In cases of this kind the agronomist has no particular responsibility. He is concerned primarily in solving the problems which call for his particular training in science. Since the better lands are already in use, most of the increased area will be made available largely by culti- vating the less favorable lands. The methods by which we shall increase the yield on lands that are 222 JOURNAL OF THE AMERICAN SOCIETY OF AGR under cultivation will be discussed under the followir 9 nree head- ings : (i) Increasing the fertility of the soil; (2)' V Her tillage methods; and (3) the improvement of crops by breed" More than half of the surface of the earth receives cipitation for the most favorable growth of crops. The it metnou of making up this deficiency is through the application of water by irrigation. Unfortunately, the supply of water for this purpose is so limited that only a fraction of the land can be served. In many cases hundreds of thousands of acres of fertile land are found adja- cent to a stream that does not contain enough water for a tenth of the land. In a case of this kind it is obvious that the volume of water and not the land area is the factor limiting production. Here the agronomist's problem lies in the direction of making the limited water produce as much as possible for each acre foot. He must call in the engineer to help in storing the water of the flood season and making it available when it can be used by crops. During the early days of irrigation no attempt at storage was made, but as the demand for water increased reservoirs were constructed, often at great cost. With the present structures, probably not more than half of the water in streams of the arid sections is fully used. The remainder runs to waste during high water or is lost through inadequate systems. One of the first steps that may be taken to in- crease food production is the construction of additional storage reser- voirs and the improvement of canals to eliminate seepage losses. Even the water that is delivered to the land falls far short of reach- ing its maximum duty. Many questions affecting the water economy of crops must still be investigated and there must be a wider applica- tion of principles of scientific irrigation before the available water will produce maximum crops. The periods when crops are most sen- sitive to water applications, the varying needs of different crops, the best methods of applying water to each type of soil, and numerous other similar questions must be investigated by the agronomist and taken fully into account before the arid regions can develop to their full fruition. It is difficult to give exact figures, but it seems probable that when all possible economies are put into operation the irrigated area of the United States can be enlarged to about four times its present size. It is largely thru the agronomist, assisted by the irrigation engineer, that this enlargement can be brought about. After all possible sources of irrigation water are fully utilized, there will be many millions of acres of arid land that can not be 5 HARRIS : THE WORLD'S FOOD SUPPLY. 223 served. \ r only possible chance for producing crops on this land is thru the >r thods of dry fanning, which means that every process is dir^' -j.rd moisture conservation. is essentially a branch of agronomy. It is based on a ,su-u a. e that will store in the soil the moisture of one or two years till it ..needed by crops. Its success depends on the selection of crops that can endure the rigors of drouth and the breeding of special drouth-resistant varieties. Probably a larger area can be added to the present productive land by the conquest of drouth than by any other means, but drouth is a relentless enemy of crop production and its successful conquest will call for all the ingenuity of students of soils and crops. Part of the preliminary work has already been done, so that one now sees grain fields where only sagebrush was found a few years- ago; but there still remain many difficulties to be overcome before all these vast areas can be made to serve the needs of man. In humid sections great tracts of land are covered, with swamps and produce no important human food. When reclaimed these lands are often exceedingly fertile. The drainage of some of the larger swamps offers rather serious engineering difficulties, but these can in most cases be overcome. The drained swamp with its peaty residue calls for special methods of management and fertilizing; but, as agronomists are seeking problems to solve, they will not be discour- aged by the difficulties encountered in changing a drained swc np into a fertile field. Somewhat related to the drainage of the swamp comes the recla- mation of alkali land, for it is largely through drainage that alkali is overcome. In all arid parts of the world the soil is likely to con- tain such an excess of soluble salts that crops can not be raised. This condition becomes more acute under irrigation. At present in the United States there are millions of acres of land that fail to produce good crops chiefly because they are impregnated with salts. In some of the western States alkali is considered to offer one of the most important and difficult problems affecting agriculture. It will be met by drainage, by special soil treatment, by breeding more resistant crops, and in other ways that agronomists may devise. The problem is now waiting; its solution will mean more food for the world. Since 1840, when Liebig explained how crops feed, great progress has been made in increasing the productivity of the soil. Before the role of mineral matter in the growth of plants was understood, all sorts of theories were advanced concerning the food used by plants 224 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. and as a result many inconsistent practices of fertilizing the soil grew up. When the real basis of plant nutrition was determined, the beginning of a rational use of fertilizers was at hand. This has re- sulted in increasing very materially the crop yields of many soils. Just how much the acre yield can be increased is uncertain. We are sure that by the proper use of fertilizers, by rotation, and by bet- ter tillage methods the present cultivated area may be made to pro- duce very much more than it is now producing, but the acre yield can not be increased indefinitely. Last year, in his presidential address before this Society, Doctor Lipman ably discussed the nitrogen problem in its relation to in- creased food production. Each element entering into commercial fertilizers might have been discussed by him with equal interest, so many are the problems surrounding the supplying of food to plants. Agronomists may be sure that they have not yet discovered every method of increasing soil fertility by the use of fertilizers. As the needs for food become more pressing, many additional discoveries will result from the researches of students of the soil. Superior tillage methods, better rotations, and many other im- provements in soil management may be expected to contribute to the increasing of the yield of the present cultivated area. So much has been done during the last few generations to improve ciops that we should hesitate before placing any limit on what may be accomplished in this respect in the future. The discovery of some of the fundamental principles of heredity has made progress much more rapid during the last few years than previously when every- thing was done by the hit-and-miss method. If no additional land could be added to the cultivated area and if there were no way to increase the fertility of the soil, considerable relief in the food situation might in time come from the breeding of crops alone ; but when this method can be taken in connection with the others, it becomes an especially valuable tool. For example, there are almost unlimited possibilities in developing crops suited to resisting drouth, soil alkali, or other unfavorable conditions in which ordinary crops can not thrive. But here too there is a limit to pos- sible improvements. From the foregoing, it is evident that the agronomist will be able to render valuable service in insuring an adequate food supply for the increasing population of the world. The question now arises as to what is his duty in the matter. Should he sit idly by as a disin- terested spectator and allow things to take their natural course, or AGRONOMIC AFFAIRS. 225 should he assume initiative and take an active part in helping to fore- stall trouble ? Will he be one who will give the ounce of prevention, or will he wait till the pound of cure is required? Probably both courses will be taken. He who is progressive, he who takes his work seriously and is anxious to use his training for the welfare of his fellows, will doubt- less take the more positive attitude and devote himself energetically to the solution of the many problems that crowd upon him. Only by profound research can these problems be solved ; but he who devotes himself honestly to seeking these solutions will find joy unspeakable and will render a lasting service to mankind. AGRONOMIC AFFAIRS. MEMBERSHIP CHANGES. The membership reported in the September-October Journal was 550. Since that report was made, 7 new members have been added and 1 has resigned, making a net membership at this time of 556. This does not take into account the fact that the dues of nearly 100 members have not been paid for 1920, which will reduce the membership by that number when these lapses are reported in the January issue. It is hoped that in the meantime a considerable part of these dues can be collected and that a large number of new mem- bers for 192 1 will be added. The names of new members and of the member who has resigned, together with such changes of address as have been reported, are as follows. New Members. Blair, A. W., Agr. Expt. Station, New Brunswick, N. J. Crocker, Leo D., High School, Jameson, Mo. Delp, H. T., High School, Boonville, Mo. Evans, A. T., Agronomy Dept., College of Agr., Brookings, S. Dak. Funk, Ernest, High School, Seymour, Mo. McQueen, Jacob, Baltic, Ohio. Mirasol, Jose J., College of Agr., Los Banos, Laguna, P. I. Member Resigned. Engle, C. C. Changes of Address. Allyx. Orr M., 618 So. First St., De Kalb, 111. Buie, T. S., Experiment Station, Florence, S. C. 226 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Clark, Chas. F., Bur. Plant Indus., U. S. Dept. Agr., Washington, D. C. Cooper, H. P., Dept. Farm Crops, Cornell Univ., Ithaca, N. Y. Cormany, Chas. E., Mich. Agr. College, East Lansing, Mich. DeYoung, Wm., Kingston, Mo. Dynes, O. W., University of Tennessee, Knoxville, Tenn. Fisher, F. A., Farm Bureau, Mt. Carmel, 111. Fuller, F. E., Henry, 111. Garber, R. J., Dept. Agron., Univ. of W. Va., Morgantown, W. Va. Haskell, S. B., Agr. Exp. Sta., Amherst, Mass. Hinde, R. R., P. O. Box 1230, Greeley, Colo. Laude, H. H., Kansas State Agr. College, Manhattan, Kans. Loomis, Howard, 205 Michigan Ave., South Haven, Mich. McMillan, S. A., Anchor, Texas. Merkle, F. G., Dept. Agron., State College, Pa. Mitchell, Jacob N., 817 Pierce Bldg., St. Louis, Mo. Moore, Harvey L., Western Union Tel. Co., Cape May, N. J. Owens, J. S., Agricultural College, Storrs, Conn. Petry, E. J., General Delivery, Brookings, S. Dak. Rothgeb, B. E., Bureau of Markets, U. S. Dept. Agr., Washington, D. C. Runk, C. R., Delaware College, Newark, Del. Slipher, John A., Natl. Lime Assn., Mather Bldg., Washington, D. C. Smith, N. S., Claresholm, Alberta, Canada. Trace, Carl F., Royster Guano Co., 321 W. Manhattan Blvd., Toledo, Ohio. Weeks, Chas. R., Secy. State Farm Bureau, Manhattan, Kans. Welch, John S., Jerome, Idaho. Wermelskirchen, Louis, 1414 University Ave., Des Moines, Iowa. Winters, N. E., 213 College Ave., Ithaca, N. Y. Woodard, John, 619 Agricultural Bldg., Urbana, 111. Wunsch, W. A., Argonia, Kans. Yost, T. F., Co. Agent Hodgeman Co., Jetmore, Kans. NOTES AND NEWS. M. A. Beeson has recalled his resignation, announced some time ago, and will remain as head of the department of agronomy of the Oklahoma college. A. M. Christensen has resigned as instructor in agronomy at the Northwest School of Agriculture, Crookston, Minn. H. P. Cooper, formerly of the Massachusetts College of Agri- culture, is now connected with the farm crops department of Cornell University. Chas. E. Cormany is now instructor in farm crops at the Michigan college. R. J. Garber, formerly of the Minnesota college, is now assistant agronomist at the West Virginia college and station. AGRONOMIC AFFAIRS. 227 REPORT OF THE SECRETARY-TREASURER. This report covers the period from November 1, 1919, to October 15, 1920. During this period 72 new members have been added to our active member- ship, 7 reinstated, 11 resigned, 3 died, and 102 allowed their membership to lapse through nonpayment of dues. As we started the year with an active membership of 473, the record shows a new loss of 37 and a total active mem- bership of 436. Thirty-five subscribers to the Journal have failed to renew their subscriptions, leaving a subscription list of 82. How to prevent this large loss each year is a problem I have been unable to solve. It is very discouraging work to build up the membership of a society by the addition of new members only to be confronted with a loss of one-fifth of the old members because' of nonpayment of dues. This loss in membership is specially unfortunate at this time when the cost of printing the Journal has been substantially raised. The question must be faced by the Society whether to cut the size of the Journal to bring it within our income or to provide additional funds. The Secretary has done a considerable amount of corresponding with advertising agencies in an effort to get them to take on the Journal on a basis that would yield a profit to the Society. Two objections are raised. The size of the Journal does not permit the use of a standard 5X8 inch plate and the mailing list is too small to appeal to advertisers. By enlarging the size of the Journal to a standard 7 X 10 inch page it would be possible to offer it in a group with agricultural college publications and derive some revenue from advertising. If any such change were made it should start with the January issue. The only other alternative in the matter of increasing funds appears to be the raising of dues to members and the rate of subscription to libraries. My personal opinion is that an increase of dues will further deplete our member- ship and that the best thing to do is to reduce the amount of printing, which amounts to over 85 per cent of our expenses, to keep within our income. At the present time the burden of supporting the Society is borne by approxi- mately one-third of the agronomists of the country. Whenever the other two- thirds take sufficient interest in the Society to join, it will be possible to main- tain the Journal at its present standard. My work in the Department of Agriculture is of such a nature that I am absent from my office sometimes for weeks and months at a time, when it is not feasible to keep up with the duties of Secretary-Treasurer. In justice to the Society I must ask to be relieved from further responsibility in that posi- tion. I wish at this time to thank the members of the Society for their many •courtesies during the three years I have served. 228 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Financial Statement from November i, 1919, to October 15, 1920. Receipts. Balance on hand from previous year $ 528.02 Dues from members : 11 members for 1919 at $2.50 $ 27.50 1 member for 1919 at i.oo a 1.00 1 member for 1919 at .50® .50 366 members for 1920 at 2.50 915.00 1 member for 1920 at 1.50 6 1.50 71 new members for 1920 at 2.50 177.50 1 new member for 1920 at 2.oo a 2.00 1 member for 1921 at 2.50 2,50 2 members for 1921 at l.oo d 2.00 2 members for 1921 r.at .50^ 1.00 1,130.50 Journal and Proceedings : 1 subscription for 1919 at 2.25 e 2.25 1 subscription for 1919 at 2.50 2.50 31 subscriptions for 1920 at 2.5o e 69.75 40 subscriptions for 1920 at 2.50 100.00 6 subscriptions for 1921 at 2.50 15.00 5 subscriptions for 1921 at 2.25 e 11.25 Sale of volumes previous to 1920 196.00 Sale of reprints 152.15 548.90 Total receipts $2,207.42 In arrears from 1919. d Advance payment. 6 Paid $1.00 in 1919. e Secured through agents. c Fifty cents still due Society. Disbursements. 1919. Nov. 4. Postage $ 9-0° Nov. 8. Printing programs Chicago meeting 6.18 Nov. 17. Postage 2.66 Nov. 18. Use of lantern, Chicago meeting 20.00 Dec. 22. Postage 15-00 Dec. 27. Postage -97 1920. Jan. 20. Maurice Joyce Engraving Co 17-95 Jan. 20. Refund on Journals and Proceedings ret'd 8.50 Jan. 31. 500 bill heads 475 Feb. 7. 400 circular letters 4- 2 5 Feb. 28. Postage .:. 13.00 Mar. 10. Maurice Joyce Engraving Co 64.97 Mar. 31. Postage .80 AGRONOMIC AFFAIRS. 229 April 9. Postage 1-44 April 28. Postage .26 May 3. Postage 1.36 May 3. New Era Printing Co 1,347.48 May 24. Postage 7-40 June 6. Postage 1 6.60 June 17. Maurice Joyce Engraving Co 10.70 June 24. Notarial fees 1.00 June 24. 1. 000 bill heads 7.00 June 24. Postage .94 July 7. Refund of overpayment for reprints 5.88 July 7. New Era Printing Co 1 18.91 Sept. 18. Postage 13.39 Sept. 18. Postage .78 Sept. 18. Notarial fees .75 Sept. 18. New Era Printing Co 356.37 Sept. 20. Maurice Joyce Engraving Co 14.50 Oct. 13. Mary R. Burr, refund for postage 2.00 Oct. 13. Mary R. Burr, clerical services 25.00 Total disbursements 2,089.79 Balance on hand •. . 117.63 $2,207.42 MINUTES OF THE THIRTEENTH ANNUAL MEETING. Springfield, Mass., October 18 and 19, 1920. First Session, Monday Afternoon, October 18. The session was called to order at 2 p. m. by President Frank S. Harris. In the absence of the Secretary, Lyman Carrier, Dr. C. R. Ball was nominated and elected Secretary pro tern. The session was devoted to a symposium on teaching agronomy, with speakers and topics as given below. It is noteworthy that every speaker was present and every paper presented in its printed sequence. Prerequisites for Agronomic Subjects, by L. E. Call, Kansas State Agr. Col- lege. Manhattan, Kans. The First College Course in Field Crops, by W. L. Slate, Jr., Connecticut Agr. College, Storrs, Conn. The Standardization of Courses in Field Crops, by J. B. Wentz, Maryland State University, College Park, Md. The Teaching of Soils in Agricultural Colleges, by W. H. Stevenson and P. E. Brown, Iowa State College, Ames, Iowa. The Teaching of Soils, by A. B. Beaumont, Massachusetts Agr. College, Amherst, Mass. The Teaching of Soils and Its Relation to Crop Subjects, by M. F. Miller, University of Missouri, Columbia, Mo. Prof. A. B. Beaumont of the Massachusetts Agricultural College extended to the Society an invitation from that institution to visit the college at 4 p.m. 23O JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. on Tuesday, with opportunity to see the intensive tobacco and onion culture in that portion of Connecticut Valley, a visit to the college and the experiment plats, dinner in the college hall and the business meeting following. On motion, the invitation was accepted. About 60 were present at the afternoon session. The following committees were appointed by the President: Nominating Committee, L. E. Call, C. JR. Ball, W. H. Stevenson. Resolutions Committee, E. O. Fippin, Robert Stewart, Geo. J. Bouyoucos. Auditing Committee, George Roberts, M. F. Miller. Second Session, Monday Evening, October 18. The second session was a joint meeting of the American Society of Agronomy and the Society for the Promotion of Agricultural Science. It was called to order at 8 p.m. by Dr. J. G. Lipman. The annual address of each of the retir- ing presidents was given as follows : The Agronomist's Part in the World's Food Supply, by Dr. Frank S. Harris, President of the American Society of Agronomy. The Future of Agricultural Science, by Dr. R. W. Thatcher, President of the Society for the Promotion of Agricultural Science. A third paper illustrated by lantern slides followed the presidential addresses. It was entitled " Methods and Results of Mosquito Extermination in New Jersey," by Dr. Thomas J. Headlee, Entomologist of the State University of New Jersey. About 100 were present. Third Session, Tuesday Morning, October 19. The session was called to order by President Harris at 9 a.m. The President asked Prof. C. G. Williams of the Ohio Agricultural Experiment Station, First Vice-President of the Society, to occupy the chair. The symposium sub- ject for the entire day was liming, and the following papers were presented, with an attendance of about 50 : The Need of Lime as Indicated by the Relative Toxicity of Acid Soil Con- ditions to Different Crops, by B. L. Hartwell, Rhode Island Agr. Expt. Sta., Kingston, R. I. The Influence of Calcium Salts on the Growth of Seedlings, by R. H. True, Univ. of Pennsylvania, Philadelphia, Pa. Liming in its Relation to Injurious Organic Compounds in the Soil, by S. D. Conner, Purdue Univ. Agr. Expt. Sta., La Fayette, Ind. The Comparative Effects of Various Forms of Lime on the Nitrogen Content of the Soil, by C. A. Mooers and W. H. Maclntire, Tennessee Agr. Expt. Sta., Knoxville, Tenn. The Influence of Liming on the Availability of Soil Potassium, Phosphorus and Sulfur, by J. K. Plummer, North Carolina Agr. Expt. Sta., Raleigh, N. C. The Nature of Soil Acidity with Regard to its Quantitative Determination, by W. H. Maclntire, Tennessee Agr. Expt. Sta., Knoxville, Tenn. * Fourth Session, Tuesday Afternoon, October 19. At 2 p.m. the session was opened by President Harris, with an attendance of more than 60 persons. Owing to the illness of Dr. T. L. Lyon he was not able to be present, and his paper, entitled " The Effect of Liming on the Com- AGRONOMIC AFFAIRS. 2 3 I position of the Drainage Water of Soils," which had been prepared and sub- mitted, was read by title only. The remainder of the papers in the liming symposium were then presented in order as follows : The Comparative Values of Burnt Lime and Limestone of Different Degrees of Fineness, by Wm. Frear, Pennsylvania Agr. Expt. Sta., State College, Pa. Comparison Between Magnesian and Non-magnesian Limestones, by A. W. Blair, Xew Jersey Agr. Expt. Sta., New Brunswick, N. J. The Value of Liming in a Crop Rotation with and without Legumes, by J. G. Lipman, New Jersey Agr. Expt. Sta., New Brunswick, N. J. Liming as Related to Farm Practice, by Frank D. Gardner, Pennsylvania Agr. Expt. Sta., State College, Pa. Owing to the plan to visit the Massachusetts Agricultural College and sur- rounding territory, beginning at 4 o'clock, the final paper of the program by Dr. W. J. Spillman and the business meeting were postponed to an evening session at Amherst. At 4 o'clock the members, to the number of 30 or 40, were taken by auto- mobile thru the Connecticut Valley for a distance of 35 miles or more, on the west side, and then brought down the east side of the valley to the Massachu- setts Agricultural College, which lies about 24 miles north of Springfield. Opportunity was given to see the intensive production of tobacco and onions, the latter at the northern end of the tobacco area. On arrival at the college, the fertilizer experiments were explained by Director Haskell and a brief in- spection was made of Stockbridge Hall, the agricultural building. The party then adjourned to the private dining room of Draper Hall, where dinner was had, followed by the annual business meeting. Annual Business Meeting, Tuesday Evening, October 10. The meeting was called to order by President Harris in the dining hall at Massachusetts Agricultural College. The minutes of the previous meeting were not present, but as they had been printed in the Journal for December, 1919 (v. 11, p. 346-348), they were approved as printed. In the absence of the Secretary-Treasurer, Lyman Carrier, the report of that officer was read by Prof. M. F. Miller. The report of the Auditing Committee was then read by Professor Miller, as follows : Report of Auditing Committee. We, the undersigned Auditing Committee, have examined the financial report and receipted vouchers for disbursements as submitted by the Secretary- Treasurer of the American Society of Agronomy and find it correct. (Signed) George Roberts, M. F. Miller. On motion the report of the Secretary-Treasurer as read was approved. In discussion of this report President Harris urged that a systematic effort be made to increase the membership to 1,000 in order that the journal of the society might be properly financed under present conditions of printing costs. On motion, the report of the Auditing Committee was approved. The report of the Nominating Committee was read by Chairman L. E. Call, and the following officers were placed in nomination : 232 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. President, Prof. Charles A. Mooers, Agronomist and Vice-Director of the Tennessee Agricultural Experiment Station. First Vice-President, Dr. S. B. Haskell, Director of the Massachusetts Agri- cultural Experiment Station. Second Vice-President, Dr. Charles B. Lipman, Soil Chemist and Bacteriolo- gist of the California Agricultural Experiment Station. Secretary-Treasurer, Dr. Percy E. Brown, Chief Soil Chemist and Bac- teriologist of the Iowa Agricultural Experiment Station. Member of the Advisory Committee on Agronomy to the National Research Council (for 5 years), Dr. Robert Stewart, Dean of Agriculture of the Uni- versity of Nevada. On motion, the report of the Nominating Committee was accepted and the Secretary instructed to cast the unanimous ballot of the Society for the nominees named, after which they were declared elected to the respective offices. The report of the Resolutions Committee was read by Chairman E. O. Fippin. Resolutions were presented and, on motion, were adopted : 1. Thanking the New England agronomists and the Massachusetts Agricul- tural College for their courtesy and hospitality. 2. Thanking the Program Committee, Messrs. Ball and Mooers, for the symposium programme prepared for the society. A third resolution asking that the Society adopt the principle of symposia in charge of qualified persons as the proper basis for future programs was dis- cussed and action thereon deferred. The reports of the following standing committees, as printed elsewhere in this issue, were then read and accepted : Committee on Terminology, read by Chairman C. V. Piper. Committee on Standardization of Field Experiments, A. T. Wiancko, Chair- man, read by the Secretary pro tern. Committee on Varietal Standardization, R. A. Oakley, Chairman, read by Prof. C. V. Piper. Advisory Committee on Agronomy to the National Research Council, read by Chairman C. V. Piper. No report was presented by the Committee on Soil Classification and Map- ping and as those engaged in soil survey activities have formed a new national organization which will officially investigate and continue this work, it was moved to discontinue this committee. The motion was carried. On motion of Prof. L. E. Call, the formation of a new committee on teach- ing agronomy was approved, the committee to be named by the newly-elected President. The third resolution presented by the Committee on Resolutions, on which action was deferred earlier in the meeting, was now brought up. After con- siderable discussion in which symposium programs for a part of the sessions, under the leadership of specially qualified persons, were warmly advocated, it was moved to approve this resolution and the motion was carried. The report of the Editor, C. W. Warburton, as printed elsewhere, was read and accepted. Mr. Warburton stated in connection to his report that he felt it necessary to resign the editorship, but as the editor is selected by the Execu- tive Committee and not elected by the Society no action was taken. On motion, a vote of thanks and appreciation was tendered to the editor for his long and faithful service. AGRONOMIC AFFAIRS. 233 A discussion of the state of the membership and finances of the Society then followed. The renewal and increase of the local sections organized previous to the war was advocated by Mr. Warburton and on his motion the dues of the Society were increased from $2.50 to $3.00 annually. The business meeting having been concluded, the program session was re- sumed and the paper by Dr. W. J. Spillman, entitled, " An Interpretation of the Results in Ohio Bulletin No. 336," was read. The Society then adjourned. Respectfully submitted, C. R. Ball, Secretary pro tern. REPORT OF COMMITTEE ON STANDARDIZATION OF FIELD EXPERIMENTS. The Committee on Standardization of Field Experiments is continuing studies and observations on methods of conducting field experiments with crops and soils and keeping in touch with the literature of the subject. It has not been possible to have a meeting of the committee this year but the mem- bers have been in touch thru correspondence. While it is felt that the Society must be careful about going on record as favoring certain systems and prac- tices, it is hoped that agreement can soon be reached on at least a few funda- . mentals which the Society can adopt and stand for. A few investigators are actively engaged in studying methods of field ex- perimentation and contributing information that will be valuable in formulat- ing standards for particular kinds of work. The committee earnestly recom- mends that all who can possibly do so should undertake some comparative study along this line because only by comparison can we hope to find what is best. Respectfully submitted, A. T. Wiancko, S. C. Salmon, A. C. Arny, Committee. REPORT OF COMMITTEE ON TERMINOLOGY. Your committeee begs to report that in spite of no new published contribu- tions, it has not been inactive, but has accumulated a great amount of material toward a completed glossary. It appears inevitable that agronomy must have a larger number of new technical terms so as to avoid ambiguity in meaning, a fault that is conspicuous in many technical papers. Perhaps we shall be forced to employ relatively as many special terms as does medicine. While there is valid objection to the introduction of new terms by the whole- sale, your committee feels that this onus is one they must bear, as individuals are timid about proposing novelties in terminology. 234 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. We ask for your continued patience, but pledge our efforts to complete the glossary in the not distant future. C. V. Piper, Chairman, C. R. Ball, H. L. Shantz. REPORT OF COMMITTEE OF VARIETAL STANDARDIZATION. At the annual meeting of the American Society of Agronomy held in Chicago in 1919, a Committee on Varietal Standardization was appointed by the Presi- dent of the Society for the purpose of considering nomenclatorial matters with regard to varieties of staple cultivated crops. The following were named by the President as members of the committee : Prof. E. F. Gaines, Washington Agricultural Experiment Station; Prof. George Stewart, Utah Agricultural Experiment Station ; Prof. J. H. Parker, Kansas Agricultural Experiment Sta- tion ; Dr. H. H. Love, Cornell University Agricultural Experiment Station; J. Allen Clark, Bureau of Plant Industry; Dr. L. H. Smith, Illinois Agri- cultural Experiment Station ; Prof. H. K. Hayes, Minnesota Agricultural Ex- periment Station ; H. G. Hastings, of Hastings Seed Co. ; A. B. Conner, Texas Agricultural Experiment Station ; and R. A. Oakley, Bureau of Plant In- dustry, Chairman. After notice of appointment was received from the President of the Society, the chairman of the committee took up correspondence with the various mem- bers for the purpose of obtaining suggestions as to the most promising lines along which it was thought practical good might be accomplished. Some very constructive suggestions were offered, and tentative plans were made by the chairman for a meeting of the committee at the annual meeting of the Society in 1920. Subsequent correspondence with the members of the committee dis- closed the fact that few, if any, would be present at the annual meeting at Springfield and, therefore, it is thought that it will be necessary to do most of the work for the coming year thru correspondence. In going over the suggestions that have been offered by members of the present committee and those offered by similar committees in the past, the Chairman is of the opinion that the greatest good that the committee can hope to accomplish will be along the line of formulating rules of nomenclature and the organization of permanent machinery whereby these rules may be con- stantly called to the attention of those working on varietal classification and to act as a court of reference in the case of controversies. It was thought at one time that field work might be planned and executed in a very helpful way, but at present it does not appear that it would be feasible to attempt work of this nature. There are a number of pieces of classification work under way and it is believed that assistance can best be rendered by this committee in the nature of cooperation and coordination in connection with existing activities. Various sets of nomenclature rules have been proposed by members of the American Society of Agronomy and other agronomic workers, and there seems to be a considerable difference of opinion with regard to the rules that should be followed. It is hoped that the Committee on Varietal Standardization will be able to reach some definite agreement as to rules of nomenclature and will also formulate workable plans for establishing and maintaining a Referee AGRONOMIC AFFAIRS. 235 Board to which all controversial questions of nomenclature may be referred. The chairman has submitted this proposal to the members of the committee, but they have had no time to reply. If they concur in these views, sub-com- mittees will be appointed in order that definite and early action may be taken. Respectfully submitted, R. A. Oakley, Chairman. REPORT OF THE ADVISORY BOARD. Your Advisory Board appointed to establish cooperative relations with the National Research Council begs to report that on March 20, 1920, it held a con- ference with the National Research Council in Washington, D. C. At this meeting were present Messrs. Call, Lipman, Mooers, Piper, and by invitation Marbut and Warburton. The whole day was devoted to a full discussion of the activities and needs of the American Society of Agronomy both as regards education and as regards research. Based on the discussions at this meeting, your Board presented a formal resolution to the Council under date of April 27, 1920. A digest of this resolution indicates its general content. 1. Detailed information regarding the American Society of Agronomy. 2. The status of the Journal of the American Society of Agronomy, which showed the need for financial assistance. A request for a contribution of $1,000 annually to aid in the publication of this journal was presented. 3. Coordination of agronomic research. — Agronomic research, particularly as regards field experimentation, requires large sums of money. The present ap- propriations are very inadequate. The investigations, both Federal and State, are, for reasons outlined, not well coordinated. The Advisory Board believes that efficient coordination can be secured best by the mutual efforts of the agronomic investigators themselves, by: (a) better planning of experiments; (b) mutual helpful criticisms of plans for securing data and of interpretation of results; (c) providing a desirable amount of duplication; (d) removing needless duplication of efforts; (e) keeping workers informed as to projects of investigation; (f) assisting to protect encroachment on research funds. In the judgment of the Advisory Board, the efficiency of agronomic research can be doubled with the present appropriations. Some assistance in such a program will doubtless be rendered by Federal and State agricultural institutions, but additional funds are vital to the success of the undertaking. It was requested that the National Research Council appropriate $10,000 annually to attain the ends desired. It is in every way desirable to promote this project as rapidly as possible, as it can be done independent of a survey of agricultural research (see proposal 4) and will do much to assist that survey if undertaken. 4. Survey of agricultural research. — Agricultural research in the United States has, like Topsy, "just grew." From the nature of its organization, all parts being more or less interallied, and with only nominal general super- vision, there is need of a broad investigation to ascertain the strong and weak points of the whole as at present developed. The Advisory Board of the American Society of Agronomy requests that this be made a special feature of the investigations of scientific research institutions now being made by the National Research Council. It would be desirable to have this investigation 236 JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. cover not agronomy alone but all agricultural research. If this proposal is adopted, the Advisory Board will be glad to furnish the Council information on many points that require consideration, not only in agronomy but in other agricultural subjects. A broad survey of the field of agricultural research will almost certainly result in findings of great importance to scientific and national welfare. 5. Suggestions as to sources of funds. — In this memorandum, a considerable list of financial and industrial institutions directly interested in the prosperity of agriculture was presented as likely sources of donations of funds for agronomic purposes. 6. Memorial urging the Council to adopt some method of publicity by which scientific men may exert a larger influence in matters that concern govern- mental policy and administration in scientific and particularly agricultural projects. The essential feature of this memorial is to bring about a better education of the public regarding the needs of agricultural research and the types of projects that need much greater support. At the present time the determination of these matters both in their inception and in their execution is mainly by nonscientific men. It is believed that a more aggressive attitude in regard to the broader problems of investigation will result appreciably in the betterment of State and governmental scientific work. Respectfully submitted, C. V. Piper, Chairman, J. G. Lipman, John W. Gilmore, L. E. Call, C. A. Mooers. REPORT OF THE EDITOR. The Journal of the American Society of Agronomy during 1920 has been hard hit by the high cost of living and of printing. The volume is the smallest which has been issued since the publication of a journal was begun in 1913. With the final number for the year, which will contain the President's address and the minutes of the annual meeting, it will contain not to exceed 248 pages as compared to 356 pages printed a year ago, and 400 pages or more in the preceding three volumes. The present volume contains 26 papers, illustrated with 8 plates and 15 text figures. These papers are by 25 authors, located in 13 States, the District of Columbia, and Guam. The principal source of con- tributions is the United States Department of Agriculture, 10 papers having originated there, while not more than two came from any other institution. The reduction in size of volume is due not to lack of contributions, but to lack of funds. In addition to the greatly advanced cost of paper and engrav- ings which has been in effect for the past two or three years, rates for com- position and press work were advanced August 1, 1919, and again May 1, 1920, making a total increase of about 65 per cent. The total cost per page now is approximately double what it was four or five years ago. While there has been a steady increase in membership during the year, we are still well below the high point reached in 1917. I can see only one way to expand the Journal and make it a really representative publication and that is by increasing both AGRONOMIC AFFAIRS. 237 the dues and the membership. I believe that an increase from $2.50 to $3.00 a year is fully justified, and that this increase can be made without ihc loss of any considerable number of members. It ought also to be possible to obtain a large number of new members in 1921, as salaries are generally becoming better adjusted to the present value of the dollar, and there are numerous agronomic workers at almost every agricultural institution who ought to be supporting the Society. Even tho the Journal is available to these workers in college and station libraries, they ought to have sufficient pride in their profession to make a small contribution each year to their organization and its publication. The Journal of the American Society of Agronomy is growing in prestige, as is evidenced by the new subscriptions which come in from time to time from foreign sources and from public libraries in the United States. These institutions not only order the current volume, but often place orders for several of the volumes published in previous years or even for a complete set. This fully justifies the policy of the executive committee in having printed a considerable larger number of copies than is required for current needs. Our income each year from the sale of back volumes, as shown by the sec- retary's records, runs fom $2co to $300, and this amount increases from year to year. The small additional cost of the copies in excess of immediate needs is much more than made up by the proceeds from future sales. Respectfully submitted, C. W. Warburton, Editor. INDEX. Page Address, Changes of, 44, 73, 114, 146, 183, 216, 225 Advisory board on agronomy to National Research Council. 148 Report of 235 Agronomists, Conference of west- ern 214 Agronomist's part in world food supply 217 Agronomic affairs, 44, 72, 114, 145, 183, 209, 225 Alfalfa species, First-generation crosses of 133 Annual meeting of Society, An- nouncement of 209 Minutes of 229 Auditing committee, Report of .. 231 Bahia grass 112 Barleys, Smooth-aWned 205 Briggs, Glen, paper on " Guam corn" Brown, Ernest B., paper on " Rela- tive yields of broken and en- tire kernels of seed corn".. Buckman, H. O., paper on " The teaching of elementary soils " Butt, N. I., see Harris, F. S. Carrier, Lyman, paper on " The history of the silo " Coefficient of selection Coefficient of yield Changes of address, 44, 73, 114, 146, 183, 216, Committee, Auditing Nominating Standardization of field ex- periments Terminology Varietal standardization.. 146, Conference of western agrono- mists Page Conference on elementary soils teaching 211 Conner, S. D., paper on " The ef- fect of zinc in soil tests with zinc and galvanized iron pots" 61 Corn crosses, Inequality of recip- rocal 185 Corn, Guam 149 Corn improvement, Selection in self-fertilized lines as a basis for 77 Corn, Planting rates in spacing for 1 Corn, seed, Formaldehyde treat- ment of 39 Yields of broken and entire.. 196 Crosses, Alfalfa 133 Corn, Inequality of reciprocal. 185 Courses in grain grading 198 Day, James W., paper on " The 149 relation, of size, shape, and number of replications of plats to probable error "... 100 196 Deceased members 146, 216 Disbursements by treasurer 228 55 Editor, Report of 236 Elementary soils teaching 55, 58 , 211 Experimental silo 69 ! Experiments, Short-time 158 175 106 Federal seed grain loans 45 168 Field experimentation, Probable error in 100 225 Field experiments, Report of com- 231 mittee on standardization of 233 232 Financial statement 228 Fippin, Elmer O., paper on " The 233 status of lime in soil im- 233 provement" 117 234 paper on " The Trufast test for sour soil " 65 2i4j Food supply, Agronomist's part in 217 238 INDEX. 239 Page Formaldehyde treatment of seed corn 39 Gaines, E. F., paper on " The in- heritance of resistance to bunt or stinking smut of wheat " 124 Garver, Samuel, see Westover, H. L. Grain grading, Course in 198 Grass, Bahia 112 Guam corn 149 Harlan, Harry V., paper on " Smooth-awned barleys " . . 205 Harris, F. S, paper on " The agronomist's part in the world's food supply" (presi- dential address) 217 and Butt, N. I., paper on " The unreliability of short-time experiments " 158 History of the silo 175 Inheritance of rust resistance in oats 23 Inheritance of smut resistance in wheat 124 Introductory courses in soils 55, 58, 211 Tones. D. F., paper on " Selection in self-fertilized lines as a basis for corn improvement " 77 Lapsed members 73 Lime, Status of, in soil improve- ment 117 Members deceased 146 lapsed 73 new.. 44, 73, 114, 145, 183, 216, 225 reinstated 114, 146, 216 resigned 114, 146, 216, 225 Membership changes, 44, 72, 114, 145. 183, 216, 225 Minutes of annual meeting 229 Mooers, C. A., paper on " Planting rates and spacing for corn under southern conditions " 1 Myers, C. H., paper on " The use of a selection coefficient " . . ic6 Page National Research Council, Ad- visory board to 75, 235 Nominating committee, report of 232 Notes and news, 44, 73, "5> 147, 183, 209, 226 Oats, Rust resistance of 23 Officers for 1921 232 Parker, John H., paper on " A pre- liminary study of the in- heritance of rust resistance in oats " 23 Planting rates for corn 1 Plat size, shape, and number, Re- lation of, to probable error 100 Probable error in field experi- ments 100 Receipts by the secretary-treasurer 228 Reciprocal corn crosses 185 Reinstated members 114, 146, 216 Report of advisory board 235 of the editor 236 of the secretary-treasurer 227 Reports of committee on auditing 231 on nominations 232 on standardization of field ex- periments 233 on terminology 233 on varietal standardization . . 234 Resigned members. . 114, 146, 216,225 Richey, Frederick D., paper on " Formaldehyde treatment of seed corn " 39 paper on " The inequality of reciprocal corn crosses ". . . . 185 Roots, Temporary, of the sor- ghums 143 Rust resistance of oats, Inheritance of 23 Scott, John M., paper on " Bahia grass" 112 Secretary-treasurer, Report of.... 227 Seed corn, Formaldehyde treat- ment of 39 24O JOURNAL OF THE AMERICAN SOCIETY OF AGRONOMY. Yields from broken and entire kernels of 196 Seed grain loans, Federal 45 Selection coefficient 106 Selection in self-fertilized lines in corn improvement 77 Short-time experiments 158 Sieglinger, John B., paper on "'Temporary roots of the sorghums " 143 Silo, Experimental 69 History of the 175 Smith, R. S., paper on " Introduc- tory courses in soils 58 Smooth-awned barleys 205 Smut resistance in wheat 124 Soil improvement, Status of lime in 117 Soils teaching, Elementary 55, 58, 211 Sorghum, Temporary roots of ... 143 Sour soil, Trufast test for 65 Spacing for corn 1 Spragg, Frank A., paper on " The coefficient of yield" 168 Standardization of field experi- ments, Report of committee on 233 Standardization of varieties, Re- port of committee on 234 Teaching soils, Conference on ele- - mentary 211 Teaching elementary soils.. 55, 58, 211 Terminology, Report of committee on 233 Trufast test for sour soil 65 Undergraduate course in grain grading 198 Varietal standardization, Commit- tee on 146 Report of committee on . 234 Waldron, L. R., paper on " First- generation crosses between two alfalfa species " 133 Warburton, C. W., paper on " Federal seed grain loans " 45 Wentz, J. B., paper on " An out- line of an undergraduate course in grain grading "... 198 Western agronomists, Conference of 214 Westover, H. L., and Garver, Samuel, paper on " A cheap and convenient experimental silo " 69 Wheat, Inheritance of bunt resist- ance in 124 World's Food supply, Agronom- ist's part in 2*7 Yield, Coefficient of 168 Zinc, Effect of, in soil tests 6i VOLUME 12 NUMBERS 8-9 JOURNAL OF THE American Society of Agronomy NOVEMBER-DECEMBER, 1920 CONTENTS The Agronomist's Part in the World's Food Supply (Presidential Address). Frank S. Harris : 217 Agronomic Affairs. Membership Changes, — Notes and News 225 Report of the Secretary-Treasurer 227 Minutes of the Thirteenth Annual Meeting 229 Reports of Committees .• 233 Report of the Editor 236 Index 238 PUBLISHED BY THE SOCIETY 41 NORTH QUEEN ST., LANCASTER, PA., and Washington, D. C. Issued December 1, 1920 Acceptance for mailing at special rate of postage provided for in section 1103, Act of October 3, 1917, authorized on June 29, 1918 JOURNAL OF THE American Society of Agronomy Issued Monthly except in June, July, and August. Editor C. W. WARBURTON Associate Editors Crops: CHARLES V. PIPER Soils: T. LYTTLETON LYON Assistant Editors Crop Production, C. A. MOOERS Soil Physics, L. E. CALL Crop Breeding, L. H. SMITH Soil Chemistry, W. P. KELLEY Crop Chemistry, R. W. THATCHER Soil Biology, J. G. LIPMAN MANUSCRIPTS Suitable articles concerned with instruction, demonstration, experimentation or research in agronomy will be accepted for publication. It is understood that articles submitted for publication have not appeared previously elsewhere and that they will not be offered for simultaneous publication in other journals without the consent of the Editor of the Journal of the American Society of Agronomy. Papers of any length, between I page and 30 or 40 pages, can be used. Personal and institutional items of agronomic interest, suitable for inclusion in " Notes and News," are solicited. To be accepted for publication, manuscripts should be original typewritten copies (not carbons) double- or triple-spaced, with wide margins. Special care should be given to the proper indication of main heads and subheads in the text, to preparation and descriptions of tables, to citations of literature and to illustrations. For fuller details see recommendations on pages 322 to 325 of volume 10 of the Journal and examples in recent issues. All illustrations desired should accompany the manuscript, should be numbered and described, and referred to in the text. Line drawings must be made in India ink and glossy velox prints of photographs are preferred for half-tones. REPRINTS Fifty reprints of each article will be furnished free. Additional copies will be sup- plied at a nominal charge. Covers with printed title page, 50 covers $i.35» and i l A cents for each additional copy. Orders for reprints and covers should he sent to the Editor immediately on receipt of proof of the article. AMERICAN SOCIETY OF AGRONOMY OFFICERS President F. S. Harris First Vice-President C. G. Williams Second Vice-President H. W. Barre Secretary-Treasurer Lyman Carrier COMMITTEES EXECUTIVE COMMITTEE Composed of the Officers of the Society - COMMITTEE ON SOIL CLASSIFICATION AND MAPPING C. F. Marbut, chairman; F. J. Alway, E. O. Fippin. J. G. Mosier, C. A. Mooers. COMMITTEE ON STANDARDIZATION OF FIELD EXPERIMENTS A. T. Wiancko, chairman; S. C. Salmon, A. C. Arny. COMMITTEE ON TERMINOLOGY Chables V. Piper, chairman; Carleton R. Ball, H. L. Shantz, Consulting Members L. C. Corbett, O. F. Cook. COMMITTEE ON VARIETAL STANDARDIZATION R. A. Oakley, chairman;- J. H. Parker, H. K. Hayes, E. F. Gaines, H. H. Love, H. S. Hastings, George Stewart, J. Allen Clark, A. B. Conner, L. H. Smith. THE AMERICAN SOCIETY OF AGRONOMY OBJECT Article II. The object of the Society shall be the increase and dissemination oi knowledge concerning soils and crops and the conditions affecting them. MEMBERSHIP Article IV. Members!. : p shall be of three kinds, active, associate and local. Active membership shall be limited to persons who are engaged in teaching agronomy or in scientific investigation in some branch of agronomy. Associate membership shall be composed of other persons interested in the object of the Society. Associate mem- bers shall be entitled to all the privileges of the Society except that of voting. Local members shall have no vote in the Society and shall not be entited to a copy of the printed proceedings without payment of an extra sum of money as provided in Article V of this Constitution. Active and associate membership may be secured by the endorsement in writing of some active member and upon approval by the President and Secretary and pay- ment of the annual dues. BY-LAWS 1. The annual dues for each active and associate member shall be $2.50, and for each local member $.50, which are due and payable on January 1 of the year for which membership is held. 2. Any member in arrears for dues for more than one year shall thereby forfeit membership, but may be restored to membership without action of the Society upon the payment of such arrears. ■ Applications for membership should be sent to the Secretary-Treasurer, preferably accompanied by remittance for dues, to save correspondence. PUBLICATIONS Proceedings. Four volumes of Proceedings have been issued, as follows: Vol. 1, cloth, 238 pp., 39 papers, 1909. Vol. 3, cloth, 286 pp., 14 papers, 19". Vol. 2, cloth, 154 pp., 16 papers, 1910. Vol. 4, cloth, 160 pp., 20 papers, 191*. Journal (continuing the Proceedings) : Vol, 5, quarterly, paper, 256 pp., 1913. Vol. 7, bimonthly, paper, 320 pp., 1915- Vol. 6, bimonthly, paper, 294 pp., 1914. Vol. 8, bimonthly, paper, 400 pp., 1916. Vols. 9, 10, and 11, monthly except June, July, and August, paper, 1917, 1918, and 1919. Price of Volumes 1 to 9, $2.00; Volumes 10 and 11, $2.50; all postpaid. Single issues, Vol. 5, 60 cents ; Vols. 6 to 8, 35 cents ; Vols. 9, 10, and 11, 30 cents. Special reduced price to members for volumes 1 to 11, inclusive. Libraries and individuals are invited to place subscriptions for the current volume and orders for previous volumes with the Secretary-Treasurer, Lyman Carrier, 41 North Queen Street, Lancaster, Pa., or U. S. Department of Agriculture, Washington, D. C.